Data mining is a process that turns raw data into actionable insights for businesses and institutions — it is a branch of analytics that finds patterns and correlations within large sets of data which can be used to predict outcomes and make decisions. 

The field of data mining isn't especially new. The term dates back to the 1980s and represents a more automated version of what has traditionally been a process of manually sifting through data for trends and patterns that has a history of more than 200 years. In modern times, data mining combines statistical methods with artificial intelligence and machine learning to rapidly assess huge volumes of data. 

The process of data mining

Data mining is sometimes said to be a misnomer because you are not actually mining for data, you are mining through data in search of patterns, trends, and anomalies that can help inform business decisions. Moreover, data mining isn't akin to a fishing expedition in which analysts review data without an overall plan; data mining is most successful when it's used with rigidly defined goals. 

The Cross-Industry Standard Process for Data Mining (CRISP-DM) is one of the leading approaches to data mining. The process can be broken down into six steps.

1. Business understanding: This is the phase in which the primary business objective is defined, along with project parameters and criteria for success. 
2. Data understanding: Analysts determine what data is needed to solve the problem identified in the business understanding. 
3. Data preparation: Frequently, the data needs to be prepared — it needs to be formatted and sanitized, fixing problems like removing corrupt data, irrelevant data, and duplicates.
4. Modeling: Algorithms are developed to identify patterns in the data. 
5. Evaluation: In this phase, analysts review the results to assess if it's addressing the objectives identified in the business understanding. The flow might need to be repeated iteratively, with the algorithm and data adjusted until the results conform to expectations. 
6. Deployment: In the final phase, the results are provided to business leaders or decision makers.  

Data mining functions and concepts

There are a lot of industry-specific terms used in relation to data mining. Here are the key concepts and functions that play a role in this process.

• Artificial intelligence (AI): AI is a computer system that can mimic some aspects of human intelligence such as planning, learning, reasoning, problem solving, and sometimes, social intelligence or creativity.
  
  
• Association rule learning: This is an analytical technique in which a system searches for relationships among variables in data. An example of this is known as market basket analysis, which Amazon uses to figure out what products are typically purchased together to make recommendations. 
• Clustering: This technique partitions data into meaningful groups or classes. It helps people and systems understand how the data should naturally or organically be structured. 
• Data analytics: The overall process of evaluating data into business intelligence.
• Data cleansing: Used in the data preparation phase, this is when raw data is put in a format suitable for analysis by eliminating data that is incorrect, incomplete, or irrelevant.
• Machine learning: This is a kind of artificial intelligence that enables computer systems to solve problems without being given an explicit algorithm. Machine learning systems can be trained or can train themselves based on examples, and the exact algorithm the software develops to solve the problem is typically unknown. 
• Regression: A common technique used to make predictions based on data. 

Uses for data mining

Data mining has a wealth of applications. It's commonly used to acquire customers, increase revenue, improve cross-selling and upselling, increase customer loyalty, detect fraud, and improving operational performance and efficiency. Here are some industries where data mining is routinely used.

• Banking: The banking industry relies on data mining to detect fraud, assess market and investment trends, and manage regulatory and compliance issues.
• Education: Educators use data mining to make predictions about student performance and develop strategies for intervening when students don't achieve the desired level.
• Manufacturing: Data mining plays an important role in detecting problems and ensuring quality on the operations floor as well as anticipating the need for equipment maintenance and forecasting customer demand. 
• Retail: This business sector is highly invested in data mining to uncover customer insights that help businesses improve sales, better target marketing campaigns, and forecast future sales trends. 

In fact, we're surrounded by real-world applications for data mining.

Amazon, for example, has an enormous amount of data about its users and what they buy, and the retailer mines that to power its recommendation engine, which provides highly targeted purchase suggestions whenever you are on the site.

Similarly, Groupon processes its enormous volume of data to continuously realign its marketing activities with customer preferences, detecting and acting on customer trends in real time. 

Join the conversation about this story »

The California Department of Motor Vehicles is looking into whether Tesla illegally misleads consumers with its claims about its "full-self driving" technology, the LA Times reported Monday and Insider confirmed.

"DMV has the matter under review," a DMV spokesperson told Insider. "The [state] regulation prohibits a company from advertising vehicles for sale or lease as autonomous unless the vehicle meets the statutory and regulatory definition of an autonomous vehicle and the company holds a deployment permit."

Tesla did not respond to a request for comment.

Tesla's FSD technology, which customers can add to their vehicles for $10,000, gives the vehicle the capability to change lanes, adjust speed, and complete some other maneuvers without assistance from the driver.

It does not make the car fully autonomous, however, according to widely accepted engineering standards, and Tesla's own website.

But the company, and specifically CEO Elon Musk, have repeatedly made ambitious promises about FSD's capabilities, only to subsequently push back the timing of new features and tout the claimed safety benefits.

Tesla has faced scrutiny over its driver-assistance features for years. But regulators and lawmakers have been taking an even closer look following Tesla allowing a small group of drivers to test a beta version of its newest FSD features.

The beta sofware has been at the center of several fatal crashes and high-profile traffic violations in recent weeks, prompting inquiries from lawmakers. Yet the company plans to roll out the software more widely even as videos posted by customers continue to show bugs that could pose major risks.

Join the conversation about this story »

NOW WATCH: A cleaning expert reveals her 3-step method for cleaning your entire home quickly

Every Tuesday, the father of venture capital, Georges Doriot, greets Two Sigma Ventures with a friendly "bonjour" and a list of potential companies to invest in. 

"Georges is a teammate," Dan Abelon, a partner at Two Sigma, said. "Constantly growing and getting smarter." 

The real Doriot, long credited with pioneering modern venture capitalism, passed away in 1987. But when Two Sigma Ventures built an artificial intelligence (AI) tool to find great companies, Abelon thought Doriot was the perfect namesake. 

"Why don't we just pick the person in history that we want most want on our team and name it after that person?" he said of his thought process. 

And so, three years ago, "Georges" was born. (Abelon affectionately refers to it as a "toddler."). Once a week, Georges sends an email with recommended companies based on an ever-growing collection of datasets, tracking things like web traffic, fundraising rounds, and mobile rankings. Abelon estimates about 12% of the fund's investments have come from Georges' suggestions, including the fund's investment in expert database NewtonX. 

Georges fits right in at Two Sigma Ventures, which primarily invests in data-driven companies. The NYC-based early stage firm is the venture arm of hedge fund Two Sigma Investments and has backed startups like Radar and sports analytic company WHOOP. While six people work on Georges full time, Abelon said they use expertise from the hundreds of data scientists across Two Sigma. 

The use of AI in venture is a rapidly growing trend. Global research firm Gartner predicted in March that 75% of venture capitalists and private equity investors will use AI to make investments by 2025 — a huge leap from the 5% that use it now, according to the Wall Street Journal. Firms like SignalFire and EQT Ventures already use data-driven tools to help determine investments, a rebuttal to the traditional VC mindset of relying on personal introductions.

"It is very much a people business, but also it's 2021," Abelon said. "It seems somewhat inconsistent for us as VCs to not be thinking about ways to use data science and software to improve our own business." 

Abelon said Georges pushes them to be better VCs, consistently suggesting companies they may not have found otherwise.  

"We want to have strong networks, but we shouldn't just rely on them," he said. "That's not fair to entrepreneurs."

So, when Georges recommends a company they wouldn't typically consider, the team makes a point to learn more and find out why it was suggested. 

"We trust Georges," Abelon said. 

The suggested startups have at times challenged the team's assumptions about venture. For example, Abelon used to be skeptical of companies that raised funds immediately after its last fundraise. However, every Tuesday, Georges' list repeatedly included companies that had fast consecutive rounds.

Instead of being a sign the company was raising too quickly, Abelon realized it was more often an indicator of rapid growth. 

Georges is constantly being updated with more datasets and fundraising information — as it matures from a "toddler" to a teen, Abelon is excited to see what else they learn.

After all, who wouldn't want the legendary Doriot on their side, even if just in AI form? 

Join the conversation about this story »

NOW WATCH: Why thoroughbred horse semen is the world's most expensive liquid

Automotive insurers are using artificial intelligence to speed up car repairs for customers, with Berkshire Hathaway-subsidiary Geico signing a deal with UK AI startup Tractable.

Tractable, founded in 2015 by a British team of computer scientists, has trained its systems on photos of damaged cars, and now promises its AI tech can give initial, accurate damage assessments to help car owners get their vehicles back from the repair shop faster.

Its technology builds on the rise of deep learning and earlier breakthroughs that now enable AI algorithms to "see" and recognize objects.

The company's system can provide insurers with initial cost estimates for car repairs at the beginning of the claims process, with Tractable relying on local data providers to estimate prices of parts and labor costs.

Alternatively, body shops can submit images of damaged vehicles along with their own cost estimates to the insurer for approval. Tractable's system analyzes the severity of the damage and either approve the estimate or flag up discrepancies to human assessors. The idea is to make the process less subjective, and make it easier for insurers and repair shops to agree on costs.

Geico, which has 17 million policyholders in the US and is the second-largest auto insurer in the country, is set to use Tractable's tech to help process the millions of claims it receives each year, the two firms said Tuesday.

Tractable and Geico have not disclosed terms of the deal. Geico CEO Todd Combs said in a statement the firm was helping it review estimates faster and more accurately.

Tractable also works with US insurer The Hartford, as well as firms in Japan, the UK, France, and Spain, and predicts that the use of artificial intelligence in insurance will become the norm. A spokeswoman for British insurance firm Ageas, one of Tractable's first clients, told Insider that its partnership with the startup had evolved from a pilot in 2017 to a fundamental part of how the firm processes claims.

The company wasn't always focused on car damage, Tractable cofounder Adrien Cohen said, initially exploring the fields of dermatology, pipe corrosion, or "any visual task that requires expertise." 

"We realized the car accidents was an amazing space to specialize in because you have hundreds of millions of claims everywhere," Cohen added. "You have a lot of data — insurance companies, body shops have taken photos of damaged cars for decades — so you can really train to reach human performance, and you can bring value to the process."

The firm plans to expand into home insurance, specifically around wind damage.

Tractable says it has trimmed its burn rate while growing topline revenue, despite COVID-19 lockdowns resulting in a fall in traffic and a subsequent decline in accidents. The company's business model is "software as a service", generating revenue from clients who sign up for subscriptions to its platform.

Cohen predicted the firm would break even within two to three years, and indicated the firm was in the process of raising funding to help its expansion. He said the firm had doubled its customer base and its staff year over year. A spokesman added the firm has more than 20 clients across 13 markets.

"We don't need to play the tech game of raising tons of money and buying your growth — our Series C was only $25 million," Cohen said. "We've really been able to contain our losses despite hypergrowth."

Join the conversation about this story »

NOW WATCH: We tested a machine that brews beer at the push of a button

Faculty, an artificial intelligence startup used by the successful Vote Leave campaign during Brexit, says it will add 400 new jobs in the coming years as it raises new cash and expands to the US.

Faculty bills itself as an "AI as a service" company that offers its tech to both governments and private businesses. It has raised £30 million ($42.4 million) from the Apax Digital Fund, part of a UK and US-focused venture capital firm.

CEO Marc Warner said the majority of the new jobs will be based in London, a hub for machine learning and AI specialists.

He added that there will be some new roles in the US, as the firm looks to expand its existing, small research and development team, sales and marketing, and eventually its delivery team. The firm will take a "phased approach" to adding new staff over the next few years, he said.

"The big expansion is likely to be in the US, ultimately it's going to be about bringing the same kinds of technology," he added.

Faculty has come in for scrutiny over its UK government ties

Faculty, founded in 2014, has proven controversial for its links to Dominic Cummings, Prime Minister Boris Johnson's former chief advisor, and its work on the Brexit campaign in 2016. Warner's brother was brought into Downing Street under Cummings as a data scientist.

Faculty has also secured 14 government contracts, totalling a value of more than £3 million, according to the Guardian. Incidentally, the Guardian's investing arm GMG Ventures is an existing backer of Faculty.

"All the contracts are on public record, we work very hard to go through the public procurement process," Warner told Insider. "All of our work has been won very legitimately."

He added that the company's work with the UK government had started before Johnson had been elected, saying its first linkup was with the Home Office under his predecessor Theresa May.

"We genuinely think it's important to work with governments and private companies and educational institutes and health systems to make this technology accessible beyond a small number of tech companies, mostly on the West coast of the US," he said.

Throughout the pandemic, the startup built a forecasting system for Britain's NHS to predict the number of cases in a hospital over a three week period. Warner said it was crucial to give hospitals confidence as to when they could start to reopen active care in certain locations as cases started to plummet.

While much of Faculty's work is kept private, Warner also explained how the company's tech had been used by a train operator to identify where vegetation had been growing onto the tracks. Such vegetation can be dangerous and damaging to trains but Faculty equipped cameras to the trains to measure what areas of the tracks trees were growing onto. Warner said it was "big important problem" that had been alleviated by the company's AI.

The company has also agreed deals with the National Crime Agency, Virgin Media, and Moonpig among others. 

Warner said that building AI systems that were worthy of trust was the company's main goal. He said there was an "incredible upside" to the widespread use of AI.

"AI isn't magic after all, it's just another application of math to understanding reality," he said.

"We cannot let the upside disappear by being too frightened, but we cannot be naive … just like we demand a safe bridge, it's no different."

Join the conversation about this story »

NOW WATCH: Why electric planes haven't taken off yet

When Raquel Urtasun joined Uber in May 2017, she walked into a battleground. 

Urtasun, a leading artificial intelligence researcher who was (and remains) a professor at the University of Toronto and a cofounder of the Vector Institute for AI, had been hired as the new R&D lead for Uber ATG, the ride-hail giant's self-driving arm, and to lead a new artificial intelligence-focused lab. 

But Uber was in turmoil. Less than three months before Urtasun joined, Waymo had launched its explosive lawsuit accusing the ride-hail giant of stealing its trade secrets. The suit centered on Anthony Levandowski, who had left Google and joined Uber the year before to lead the autonomy effort that CEO Travis Kalanick considered vital to his company's long-term survival. 

Uber had sidelined (and would soon fire) Levandowski, who was ultimately convicted on a criminal charge of trade secret theft (only to be pardoned by an outgoing President Trump). Kalanick was himself under fire, as a series of scandals exposed Uber's toxic, often misogynistic corporate culture, and would resign in June. 

Considering the chaos, Urtasun appeared to be a brilliant hire for ATG. Not only could her pioneering work in artificial intelligence be untouched by any claims of intellectual property theft, it could seriously propel the company's self-driving ambitions.

That didn't quite work out. While she was respected and well-liked, according to two former Uber engineers who interacted with Urtasun during her tenure, her impact on Uber's self-driving program was limited. 

Part of the problem was that Urtasun's novel machine learning algorithms weren't easily plugged into a software stack that was several years old when she arrived. "She didn't have a big impact," one of the former engineers told Insider. "To her frustration and that of others. She clashed with the team because you can't just hand off an algorithm to a roboticist." 

In response, Urtasun said she "had a significant impact on ATG's software stack," and that her time leading R&D at the ride-hail giant taught her what worked — and didn't — in using advanced AI in self-driving cars. "From my perspective, AI needs to be at the center of the solution in order for the technology to scale," she said. "It can't be peripheral or plugged into an old software stack."

The idea that Urtasun — or anyone — could rehab Uber's self-driving program fully fell apart in March 2018, when one of its self-driving cars killed a woman crossing the street in Arizona, in part because it failed to identify her as a pedestrian. Uber essentially iced the effort, letting it limp along with minimal on-street testing until December 2020, when it offloaded it to Aurora. Aurora did not make an offer to Urtasun and her team, TechCrunch reported. 

A fresh start

Now, Urtasun is returning to the self-driving game with a startup of her own, free of baggage and loaded with cash. She is the founder and CEO of Waabi, the Toronto-based startup that emerged from stealth Tuesday morning, announcing it had raised $83.5 million in Series A funding. Khosla Ventures led the round, with participation from Uber, Aurora, 8VC, Radical Ventures, Omers Ventures, BDC, and AI bigwigs Geoffrey Hinton and Fei-Fei Li. 

The startup will focus on autonomous trucks, where a clear commercial use case and a relatively simple driving environment make for an easier problem than piloting a robo-taxi through a city center. And rather than build its own vehicle from the ground up, Urtasun said Waabi will be "very partnership friendly," eager to team up with existing manufacturers, hardware providers, and so on. 

Urtasun launched her own company because in a field dominated by a handful of players — Waymo, Cruise, Argo, and Aurora chief among them — she sees a whole lot of same. 

"What's happening in the industry is that there is really one way of solving this problem," she told Insider, noting that most of the leaders of the field share a pedigree, having started off on the early Google Chauffeur team that sprang out of DARPA's Grand and Urban Challenges. Those competitions, held between 2004 and 2007, sparked today's self-driving industry and did much to set the recipe for how an autonomous vehicle is made. After more than a decade of work and billions spent, commercial self-driving deployments are rare and small. 

"The world is trying to solve this very complex task in the same way," she said. "So there is a need for a diversity of solutions."

Urtasun's pitch is that by leaning heavily on artificial intelligence, namely deep learning, Waabi can move much faster than her established competitors, with many fewer engineers and much less investment. "If you want to do that wholesale, the best route is to start a new company," she said. 

New blood

Still, the idea of launching an autonomy-focused startup feels very 2016. Over the past few years, as the sheer difficulty of the problem has become clear, investment in the space has cooled and consolidation has been the order. The prevailing wisdom is that making a car drive itself takes a certain amount of brute force. 

"I can't think of another example of a situation where you have to get to the point where you have literally thousands of engineers, billions of dollars of capital, and no product yet," Cruise CEO Dan Ammann said in a 2019 interview. "This is not a problem you can solve with 50 or 100 engineers."

Urtasun is looking for the middle ground. She'll use her new funding to grow her current team of about 40 (many of them holdovers from her Uber days), but she isn't building a juggernaut. And that clearly has its appeal. 

While some investors believe the likes of Waymo and Cruise will dominate, "others believe there's a need for new blood, smart people taking a new approach," said Shahin Farschi, a partner at Lux Capital who was an early investor in Zoox. Waabi's impressive Series A makes that clear. 

But AV trucking is an already crowded space. If Waabi wants to form partnerships, Farschi said,  "the question becomes, how do you integrate with those teams?" 

And that's a problem Waabi will have to solve with more than artificial intelligence.

Join the conversation about this story »

NOW WATCH: Why we still haven't been able to find aliens

An exchange-traded fund that uses artificial intelligence in identifying the most promising US equities has blasted past the S&P 500 in returns for the past month — without any meme stocks in its holdings, analysis by DataTrek showed.

The AIEQ ETF's performance for the past month has been especially strong, and as of Monday, the 1-month return for the fund was 10.6%, compared with 2% for the S&P 500. The 1-year return was 50.1%, while it was 39.9% for the index, DataTrek found.

Year-to-date, though, the two are much closer, at 13.6% for the ETF and 13.3% for the US equity index. And while the AIEQ has been logging a solid performance overall, its 3-month return was 0.5%, much lower than the S&P 500's 7.2%. 

The ETF regained strength and caught up with the index over the last month. It did this by adding companies with better performances to its holdings, DataTrek said. "In that respect, AIEQ's process looks a lot like a human manager, searching for momentum names that fit its investment process." the financial research firm said. 

Despite this, DataTrek analysis showed the ETF was free of meme stocks, which can rise and fall in quick succession, significantly and quickly impacting returns. 

Meme stocks such as GameStop and AMC have sent rumbles through the market in recent months, after retail investors banded together to buy the shares, driving prices up and causing short squeezes. As this caused stocks to be overvalued, those who invested at the highs or during the run-up were left vulnerable to big losses in the longer term. 

Instead, AIEQ shifted its top 10 holdings. It kept Alphabet, 10X Genomics, Costar Group, Tesla and Square in the mix, albeit with weighting adjustments, and added MongoDB, DexCom, Appian, Carvana and Autozone. DocuSign and Yeti Holdings were among the stocks moved out of the top 10. 

The fund also spread out its investments further. Its top 10 holdings now account for 28% of the portfolio, versus 40% previously, in an effort to catch the momentum of a wider range of stocks, DataTrek said. It currently manages assets worth $155.6 million and trades on the NYSE ARCA exchange. 

AIEQ is powered by IBM's Watson supercomputing, which enables the actively-managed ETF to use artificial intelligence when picking stocks. It analyzes company, technical, macro and market data from news, social media, industry analysis, and financial statements, according to its profile on the website of the ETF Managers Trust, which runs the fund.

Join the conversation about this story »

NOW WATCH: What happens when you drink too much water

Chinese fighter pilots have been battling aircraft piloted by artificial intelligence in simulated dogfights to boost pilot combat skills, Chinese media reported.

Fang Guoyu, a People's Liberation Army Air Force brigade flight team leader and recognized fighter ace, was recently "shot down" by an AI adversary in an air-to-air combat simulation, according to China's PLA Daily, the official newspaper of the Chinese military.

He said that early in the training, it was easy to defeat the AI adversary. But with each round of combat, the AI reportedly learned from its human opponent. After one fight that Fang won with a bit of skillful flying, the AI came back and used the same tactics against him, defeating the human pilot.

"It's like a digital 'Golden Helmet' pilot that excels at learning, assimilating, reviewing, and researching," Fang said, referring to the elite pilots who emerge victorious in the "Golden Helmet" air combat contests. "The move with which you defeated it today will be in its hands tomorrow."

Du Jianfeng, the brigade commander, told the PLA newspaper that AI is increasingly being incorporated into training.

It "is skilled at handling the aircraft and makes flawless tactical decisions," he said, characterizing the AI adversary as a useful tool for "sharpening the sword" because it forces the Chinese pilots to get more creative.

'Sharpening the sword'

China is striving to build a modern military with the ability to fight and win wars by the middle of this century, and it has made progress in recent years in advancing its air combat element, even developing a fifth-generation stealth fighter.

But far more challenging and time consuming than closing the technology gap is cultivating the critical knowledge and experience required to effectively operate a modern fighting force.

Chinese media did not really offer specifics on the simulator, which was developed by the military in cooperation with research institutes, so there are still some questions about whether or not the AI adversary provides sufficiently realistic training necessary to prepare pilots to dogfight manned aircraft.

"If it does, that's pretty good," retired US Navy Cmdr. Guy Snodgrass, a former TOPGUN instructor and an artificial intelligence expert, told Insider.

"If it doesn't," he continued, "you're really just training human operators to fight AI, and that is probably not what they are going to be going up against" since there are currently no autonomous AI-driven fighter aircraft they would need to be prepared to fight.

"There could be a divergence between real capability in a dogfight or aerial battle versus what the AI is presenting," he said. If that's the case, it could be wasted effort.

If it is a high-fidelity training simulator, though, it potentially lowers the cost of the air combat training because "you're able to get that training at a price point that's much lower than actually putting real planes in the air," Snodgrass said. It also lowers the risk.

Chinese leader Xi Jinping has repeatedly stressed the need for realistic combat training, including simulations, to help the Chinese military overcome their lack of combat experience, but it is not clear to what extent his agenda has been implemented with training simulators like what PLAAF pilots have been using.

'The AI is learning and it's getting better'

Regardless of whether the pilots are learning anything valuable, Fang Guoyu's recollection of his engagements with his AI adversary demonstrates that the AI is.

"AI requires feedback," Snodgrass said. "And that's exactly the kind of pathway you'd want to take, to use this to help train your pilots, but because your pilots are fighting against it, the AI is learning and it's getting better."

A next step, he explained, could then be to say, "This has performed very well in a virtual environment. Let's put this into a manned fighter."

China has invested heavily in AI research, and, like the US, it has been considering ways to incorporate AI — which can process information quickly and gain years of experience in a very short time — into the cockpits of its planes.

Yang Wei, chief designer for the J-20, China's first fifth-generation stealth fighter, said last year that the next generation of fighter could feature AI systems able to assist pilots with decisions to increase their overall effectiveness in combat, the state-affiliated Global Times reported.

The US Air Force has expressed similar ideas. Steven Rogers, a senior scientist at the US Air Force Research Laboratory, told Inside Defense in 2018 that ace pilots have thousands of hours of experience. Then he asked, "What happens if I can augment their ability with a system that can have literally millions of hours of training time?"

Snodgrass explained that there are a number of different ways AI could be used to augment the capabilities of a pilot.

For instance, artificial intelligence could be used to monitor aircraft systems to reduce task saturation, especially for single-pilot aircraft, collect battlefield information, and handle target discrimination and prioritization. AI could even potentially chart out flight paths to minimize detection through electromagnetic spectrum analysis.

The US is currently pursuing several lines of effort exploring the possibilities of AI technology.

In a big event last summer, the Defense Advanced Research Projects Agency (DARPA) put an AI algorithm up against an experienced human pilot in a "simulated within-visual-range air combat" situation.

The artificial intelligence, which had already defeated other AI "pilots" in simulated dogfights and collected years of experience in a matter of months, achieved a flawless victory, winning five straight matches without the human, a US Air Force F-16 pilot, ever scoring a hit.

The point of the simulated air-to-air combat scenario was to move DARPA's Air Combat Evolution program forward.

The agency said previously that it envisions "a future in which AI handles the split-second maneuvering during within-visual-range dogfights, keeping pilots safer and more effective as they orchestrate large numbers of unmanned systems into a web of overwhelming combat effects."

It is not clear how long it would take to realize the agency's vision for the future, but Snodgrass previously told Insider that he "would never bet against technological progress," especially considering "all the advancements that have occurred in the last decade, in the last hundred years."

Join the conversation about this story »

NOW WATCH: One bite from this tick could ruin red meat for you — or even kill you

Tractable says it is the UK's first computer vision unicorn, after raising a $60 million in fresh funding at a valuation above $1 billion.

The round was led by existing backers Insight Partners and Georgian.

Company president Adrien Cohen said the firm had grown revenue to eight figures, thanks to the startup landing major insurers as clients in the last year.

The London startup, founded in 2014, has primarily applied computer vision capabilities to assess car damage after an accident. It partners with insurers to help make an initial assessment and estimate repair costs, something which can reduce the time a car spends in the body shop. The firm has trained its algorithms on scores of photos of damaged cars, and claims the system is as accurate as a human.

Tractable was cofounded by computer scientist and former hedge fund quant Alex Dalyac, machine learning specialist Razvan Ranca, and ex-Lazada exec Adrien Cohen.

"Reaching this milestone is not important, per se, but it's what it says about the impact and scale of our technology, the validation of reaching this scale," Cohen said of the firm's unicorn status. The startup counts around 20 insurance clients across the US, Europe, and Asia, including Berkshire Hathaway affiliate Geico.

Though initially specializing in auto repair assessments and estimates, the firm is now expanding into analyzing property damage and even car purchasing.

"We're going to go deeper, we think our AI can deal with cases where you want to inspect a vehicle's condition, not just in an accident, so when you purchase, sell, or lease" said Cohen. "All these events, where you can accelerate the process by understanding the vehicle condition from photos."

In theory, the platform could partner with a used-car platform like Cazoo to assess the condition of a car placed for sale. Cohen said auto rental firms and auto manufacturers are also potential clients.

Asked about revenue growth, Cohen said the firm was privately held and would not reveal specific numbers. "It's an 8-figure revenue [number], with 600% growth in the past 24 months," he said, adding that the firm had only raised $55 million in outside capital prior to the new round.

Computer vision startups and the route to commercialization

Tractable is one of a wave of startups benefiting from the maturation of computer vision. According to this year's edition of the annual AI Index, collated with Stanford University, computer vision is becoming increasingly "industrialized."

Alessio Petrozziello, machine learning research engineer at London data extraction startup Evolution AI, says that more broadly computer vision has some hurdles to clear before it goes fully mainstream.

"There's certainly a push to commercialize these models, but it's been clear they are not at the level where you can [fully] rely on them," he said. "For example for a self-driving car, it can't make any mistakes, certainly no more than a human." Apart from accuracy, he added, there's the issue of responsibility. "You use a model, and the model makes a mistake, who's responsible? There isn't a clear-cut answer."

Eleonora Ferrero, director of operations at Evolution AI, added that success for startups like Tractable was as much about execution as the fundamental computer vision tech.

"Their go-to-market was partnerships with key insurance companies that provided data, it's an advantage," she said, adding that Tractable had been smart to identify something that insurers sought — increased operational efficiency.

Karen Burns, founder of computer vision platform Fyma, said adoption depended on clients being ready for the tech. Fyma's platform, trained on anonymized data, analyzes what's going on in a physical space — whether that's a firm tracking the movements of its autonomous robots for safety; or a retailer measuring footfall.

"Before you can adopt AI, you have to go through a big transformation," she said.

Tractable's Cohen agreed, saying that the firm had relied on the quality of its AI development but also selling the usefulness of AI to enterprise clients. "A big playbook we've cracked is how to deploy and capture the value of artificial intelligence in an enterprise context," he said. "This is very challenging, and something we've had to figure out."

Join the conversation about this story »

NOW WATCH: How long humans could survive in space without a spacesuit

As people do more of their shopping online, getting people to buy once is not enough for marketers. Getting them to stick around is the holy grail.

Former Paypal and Facebook product and data analytics manager Emad Hasan just raised $8 million in Series A funding for his startup Retina, which he says helps brands like Dollar Shave Club and Madison Reed acquire and retain customers.

The round was led by Alpha Intelligence Capital and Vertical Venture Partners. Retina plans to use the funding to boost hiring and expand integrations with companies including Facebook, Google, Shopify, Segment, and Experian. Past investors include Comcast Ventures.

Retina uses artificial intelligence, machine learning, and data analytics to predict customers' lifetime value — the measure of a customer's total worth over the course of their relationship. 

It claims to do this with a proprietary algorithm that looks at companies' order history and customer attributes and builds lookalike audiences to help them boost sales. 

Below is Retina's fundraising deck.

Founded in 2017, Retina is a predictive customer intelligence startup that says it helps brands with targeting, ad relevance, and customer loyalty.

Its pitch is that instead of looking at customer acquisition costs, brands should look at customers' lifetime value.

Retina co-founder and CEO Emad Hasan said Facebook and Google are great customer acquisition channels but take months to share return on investment intel.

Retina says brands should focus on customer lifetime value, which it calls a $500 million-plus revenue opportunity.

Retina claims its platform can bring insights to marketers even before a customer makes their first purchase.

It claims to do this with an algorithm that creates consumer profiles based on brands' data.

Brands use Retina to improve their ad campaigns and test new products. The company plans to expand to customer service to let brands prioritize high-value customers.

Retina says its clients include Nestle, Madison Reed, Dollar Shave Club, and Brickell Men's Products.

It says one client used the platform to improve its return on ad spending in real-time.

Retina said another client used the platform to test which product features its high-value customers sought, and tweaked its advertising accordingly.

Hasan's cofounder is Michael Greenberg, an entrepreneur that founded Scale Funder and an academic at UCLA.

The company plans to use the funding to hire, create a self-service product, and work with companies including Shopify+, Segment, Experian, Facebook, and Google.

Early in the pandemic, the board of AI startup Hypersonix asked if the company was planning to revise its projections for the year ahead. The company said no, there was no need: It was already running pretty efficiently, its leadership tol the board, and it had plenty of venture capital cash in the bank just in case.

"We changed processes and we tried to do business, but we charged forward," Hypersonix founder and CEO Prem Kiran told Insider. "We've been smart in pivoting where we need to pivot."

Kiran says Hypersonix saw "tremendous" growth in the past year as many companies started relying more on data as they brought their businesses online. For example, many of its clients are from the grocery and retail sectors, which saw an increase in demand as online shopping surged, which in turn helped bring in more cash to Hypersonix's business. 

This growth helped Hypersonix raise a $35 million Series B led by B Capital Group, the company announced Wednesday. Its valuation also quadrupled to $200 million in just over a year since its Series A.

Hypersonix plans to use the funding to scale by expanding to more industries and locations and building new partnerships. It also plans to invest in its product and technology. 

"As we get into the next level of scale, we want to have more and more automation to the product," Kiran said. "The more automated the product is, the faster the product is. It becomes the ultimate advantage."

Hypersonix could compete better against 'traditional big boys' like Oracle and SAP during the pandemic

Kiran, an 11-year veteran of SAP before becoming an entrepreneur, founded Hypersonix in 2018 out of his "own frustrations." At previous companies, every time he and other executives tried to make a decision like a price change, he had to go through a lengthy process and had to consult multiple people and analysts. Even for basic questions, it could take two to three days to turn around. In some cases, it could take weeks. 

Kiran decided to build a system that tells people the pros and cons of decisions they want to make to help them weigh their options. It can help users get all the data they need to make a decision in minutes, or even seconds. 

"The amount of time it takes to get an answer to a simple question like that, it was astronomical," Kiran said. "I was shocked."

Hypersonix competes with tech giants like Oracle and Kiran's former employer SAP. But as a smaller startup, Hypersonix has speed and agility on its side, Kiran said. 

"They cannot store and process and compute and stitch data and leverage data in the way we're able to do it," Kiran said. "Hence what we're seeing is time to value. We're able to get our customers up and running faster."

Transitioning to remote work during the pandemic also helped Hypersonix compete with the "traditional big boys," Kiran said. That's because all companies are having meetings with their large clients on Zoom or other videoconferencing software, which puts them all at a "level playing field," especially since a startup like Hypersonix may not have as large of a salesforce.

"No one had a different benefit over us," Kiran said. "It gave us a fair shot at large clients."

Join the conversation about this story »

NOW WATCH: What it's like in the death zone of Everest, K2, and other mountains

Artificial intelligence-powered investing is not garnering as much attention as it used to, but it just might be making a comeback in the age of meme stocks and Reddit traders. 

Just look at the Van Eck Vectors Social Sentiment ETF (BUZZ). The fund, famously backed by Barstool Sports' Dave Portnoy, uses artificial intelligence technology to screen for stocks with positive investor sentiment. It has amassed about $250 million in assets and was up 12.9% since its launch four months ago. 

EquBot, another ETF provider most well-known for launching the first artificial intelligence-powered equity ETF, also offers a family of funds poised to take advantage of the meme investing trend. The firm leverages IBM Watson-powered artificial intelligence to gather millions of data points every day to build on multiple predictive models. 

"Utilizing IBM Watson and our own proprietary algorithms, our model is focused on what and who is moving the markets and how fast, so it helps us understand better not just what to invest in but when," Chris Natividad, chief investment officer of EquBot said in a recent interview. 

The timing is important. Despite the meteoric rise of AMC and GameStop this year, their upward trajectories are punctuated by periods of sharp declines. That's why EquBot's model averted GameStop when it was first surging in January. 

During that initial spike, the model captured a huge spike in keywords sentiment through various social media channels, but it was also aware that GameStop is predominantly driven by other metrics such as in-store sales, the supply and demand dynamic for its products, and different foot traffic metrics, Natividad said. 

"What happens is there's a dynamic waiting because the system sees the market price is really dislocated from the traditional set of signals that have correlated to the GameStop price, and the confidence in the predictive financial model decreases," he added. "Ultimately, it's not going to take on that position, which was the right move as far as the timing when the system was ingesting that information."

Because the two funds offered by EquBot — AI Powered International Equity ETF (AIIQ) and AI Powered Equity ETF (AIEQ) — are both actively managed, their portfolio turnover rate is on the higher end typically at around 2% per day. However, depending on the day and market dynamics, sometimes the data-dependent models would not signal any trades, Natividad said.

With a bird's eye view of market sentiment and trends, Natividad recently shared five global stocks showing strong growth signals, as predicted by the AI model. The five stocks, along with their tickers, market caps, and commentaries, are listed below. (Natividad's stock commentaries are as of June 30.)

1. Atlassian Corporation

Ticker: TEAM

Market cap: $64.88 billion 

Commentary: "Global companies are adapting to the evolving remote workplace environment. Atlassian has created a growing productivity software suite that is changing how teams work virtually. Shares are up north of 15% year-to-date and closed a couple dollars short of its 52-week high at 267.96."

Source: EquBot

2. Shopify

Ticker: SHOP

Market cap: $181.82 billion 

Commentary: "Often dubbed the "Canadian Amazon" Shopify seemingly achieves higher ratings for smaller ecommerce players.  Shopify was positioned in the sweet spot during global pandemic lockdowns and its stock has increased over 65% in the past year. Ecommerce data continues to grow and there is an increased level of stickiness to consumer behavior in repeated electronic transactions free of complications."

Source: EquBot

3. Crispr Therapeutics

Ticker: CRSP

Market cap: $12.28 billion 

Commentary: "Technological advancements proximal to healthcare and pharmaceutical sectors have elevated levels of interest as a result of the significant feats accomplished throughout the COVID pandemic. CRISPR is a gene editing leader with Nobel prize-winning ties. Transformative DNA healthcare applications can have profound impact. Shares are down year to date about 18% and last traded at 129.20 but the EquBot system is showing strong positive long term signals related to aggregate market perception and management team influence."

Source: EquBot

4. Spotify Technologies

Ticker: SPOT

Market cap: $51.91 billion

Commentary: "Spotify is the leading streaming music provider and the platform user counts continued to grow throughout the outbreak of COVID-19. Spotify has been in competition with tech behemoths for years and is still leading the charge. Data on streaming music as well as podcast participation for Spotify continue to be encouraging."

Source: EquBot

5. Wix.com

Ticker: WIX

Market cap: $16.28 billion 

Commentary: "The ecommerce software company from Israel providing cloud-based web development services is yet another tech example that has found itself positioned for continued growth as global economies recover. The platform experience seems to be well received across a variety of consumer mediums and shows elevated sentiment scores from business partners and equally constructive analyst ratings from Wall St. We did see WIX stock prices pull back as a result of proximal geopolitical conflict earlier in the year but it is seemingly back to marching up another strong year from a stock price perspective."

Source: EquBot

Interest in prefab tiny homes skyrocketed during COVID-19.

Now one Singapore-based company is looking to capitalize on this trend by introducing its artificial-intelligence-powered tiny homes to the UK.

Meet Nestron, the brains behind several wildly popular tiny homes that have since become a "worldwide phenomenon," Choco Toh of Nestron's marketing team told Insider in December.

Source: Insider

Its tiny homes were such a hit, Nestron's website crashed for a while, likely due to an influx in visits and "extremely overwhelming" popularity, Toh said.

Source: Insider

To expand its reach, Nestron is now in the process of preparing its debut in Northampton, England, a little more than 65 miles from London.

Toh says Nestron will close about 10 deals before the homes actually debut in Europe …

… but estimates that by the end of the year, it'll sell more than 100 units in the UK.

"We believe with the increase in marketing activities upon our debut, there are nearly 100,000 potential users in the UK, which will bring explosive and continuous growth to our local distributors," Toh said in an email.

Like other companies that ship products internationally, Nestron has struggled to move its tiny homes in the face of jammed ports and shipping delays.

But before we dive into how the company is overcoming these issues, let's take a look at the two futuristic tiny homes that will debut in the UK: the $34,000 to $52,000 Cube One and the $59,000 to $77,000 Cube Two.

These prices vary widely due to a list of possible extra add-ons, such as solar panels, heated floors, and additional smart appliances.

The Cube One is more popular with solo occupants, while the larger Cube Two has been a hit with families, couples, and as a backyard unit.

Nestron debuted both units well before its UK plans but has since made sizing changes ahead of its overseas delivery: The Cube One's size was boosted about 16.2 square feet, while the Cube Two was expanded by about 25 square feet.

Let's take a closer look at the Cube One, which stands at about 156 square feet.

That square footage holds the living room, bedroom, bathroom, and kitchen space (which comes with cabinets, a sink, and a stovetop, according to renderings of the unit).

As with any typical home, the living room has a dining table and sofa, while the bedroom has a side table, closet, and of course, a bed.

Moving toward the bathroom, the tiny Cube One comes with a shower, towel rack, and sink, all in one space.

The little living unit also has built-in amenities like lights, storage units, electric blinds, and a speaker.

There's even room for a modern-day must-have: air-conditioning units.

Now let's take a look at the larger Cube Two, which can accommodate three to four people with its two beds, which sit on opposite ends of the tiny home.

As with its smaller sibling, the almost 280-square-foot Cube Two has a living room, two beds, a kitchen, and a bathroom, all with the same furnishings as the Cube One.

But the dining table in the Cube Two is noticeably larger, and there's a skylight for added natural light and stargazing.

Both models come insulated and have smart-home capabilities using Nestron's "Canny," an artificial-intelligence system.

Canny can complete tasks like brewing your morning coffee or automatically adjusting your seat heights.

Everything is "smart" these days, which means the Cube One and Two can also come with motion-sensing lights and smart mirrors and toilets.

You might be wondering how Nestron plans to move its Cube One and Two tiny homes overseas in one piece.

According to Toh, Nestron has had a "solid foundation built in the industry ... allowing it to have a good relationship with experienced and professional forwarding partners."

Despite this foundation, as with other companies, Nestron has experienced delays related to the global supply-chain jam, specifically congested ports in the UK.

As a result, the company's forwarding charges were triple what it initially expected, Toh said.

But instead of charging its clients extra money for immediate shipping, Nestron decided it would pause shipping until costs came down.

To bypass these congestion issues, Nestron also decided to reroute its original plan to ship straight to the UK.

Nestron "decided to travel over to Antwerp, Belgium, and then land in the UK," Toh said. "This way, by the time we reach the UK port, the congestion would've been clear."

Despite this detour, shipping costs were still higher than expected, in part because the company and its distributors still wanted to make the debut timeline.

"Since the demands are growing and people want to experience touch and feel with Nestron, we took the chance and sent the units off earlier this month," Toh said, expecting them to arrive in late July or early August.

To aid in the transportation process, the tiny homes have built-in retractable hooks to help make them compatible with cranes.

The homes' structures are also stable enough to withstand the stress of moving, Toh said.

And all the little living units are also packaged in waterproof fabric to both avoid rusting and also to allow for easy inspection.

Being in the UK will allow potential consumers to "engage with Nestron units directly," Toh said. "The experience will definitely influence the market interest and purchase power."

Incumbent financial institutions often have notoriously slow, clunky technology — but that's exactly what lured Manuela Veloso, Carnegie Mellon University's head of machine learning, to JPMorgan Chase three years ago. 

"JPMorgan has been around for a long time. Unlike some technology companies like Amazon or Google, JPMorgan wasn't born digital and that was one of the challenges that attracted me: How do you bring AI to these non-digital services?" Veloso told Insider.

Veloso is the head of AI research at JPMorgan, currently on leave from her role at the university. Her work, unlike many tech teams on Wall Street, is not driven by business needs. Instead, Veloso mulls the infinite what-ifs. 

What if Chase gained 2 million online-only users tomorrow? What if the Federal Reserve aggressively hiked the interest rate this week? And what are the ripple effects of those what-ifs? 

Exploration is key to Veloso's research. It's one of her goals as JPM's lead AI researcher to "make people embrace, grow their hearts, for the use of AI," even if it doesn't necessarily lead to tech deployment or a business tool. 

Veloso's at the helm of a team of nearly 60 academics whose expertise span mathematics, cryptography, electrical engineering, and machine learning, to name a few.

She outlined the seven main challenges her team is trying to solve with AI.

Financial crime

JPMorgan is researching how AI can help eradicate financial crime, which is one of the most frustrating problems Veloso said she has encountered since joining the bank.

"I didn't anticipate financial crime would be such a major focus area for my group," she said.

Identity theft will cost US firms $721 billion in 2021, up 42% from 2019's $502 billion loss, according to a March report from Aite Group, a financial research and advisory firm. In that timeframe, nearly half (47%) of American respondents reported being victims of some kind of identity theft, like account takeover or application fraud.

One way to fight financial crime is exploring how data and AI can improve fraud and anomaly detection, she said. The bank is working to expand its detection beyond single events and instead train AI to identify behavioral patterns, or sequences of events, that are indicative of financial crime. Graphs are used to model these behavioral networks, she added. 

Large economic systems

Veloso's team uses AI to predict the impact of interactions between the various players within large economic systems, often referred to as multi-agent systems. Examples include supply chains or when multiple parties are involved in a trade or exchange. 

A key part of this — and an underplayed application that people have yet to take advantage of — is the use of AI-powered simulations to understand and test the ripple effects of those interactions, Veloso said. But to understand that, researchers have to go beyond historical data, she added.

Real data, while valuable, doesn't necessarily allow the user to expand into different possibilities. Veloso likened deploying AI to historical data to a copy machine. Whatever happens can be reproduced, "but it doesn't let you go beyond the cycle. Simulations enable you to explore a much larger space than what the real data tells you," she said.

Datamanagement

Despite the limitations historical data can have when deploying AI, data is still embedded into JPMorgan's framework. Veloso's team is exploring how to use AI for data management, like sharing data safely between multiple divisions within a bank or externally, she said. 

The bank is using AI to fetch public information and help with the controls of who has access and sharing permissions, Veloso said. There's a lot of emphasis on the infrastructure, specifically about having that data available on the cloud, she added.

The team is also deepening its understanding of how cryptography and synthetic data can help with data privacy and protection, she added.

Employee empowerment

Another main goal of the research team is figuring out how AI can improve how employees work.

That includes using internal chatbots to help call center agents siphon through bank resources, using AI methods to classify emails, and automating the generation of reports and PowerPoints, Veloso added.

Client experience

Veloso's work is also focused on the customer experience. That includes how quickly employees access data for customers, and using data to streamline the customer experience.

For example, information required for know-your-customer regulations often already exists on JPMorgan documents or publicly-available ones. However, mapping the existing information to a given data field is a manual process, Veloso said.

Now, JPMorgan is automating that with natural-language processing and presenting the customer with a form pre-filled by AI that they just have to check is correct, she said.

Regulation and compliance 

As a global bank, JPMorgan has thousands of obligations and rules it must understand, and know which regulations apply to which jurisdictions and businesses. 

If a banker talks on the phone with a client from Singapore, or they write a contract with a client in Italy, understanding what laws govern what they do is critical.

It's a complex system to navigate, Veloso said. The bank is analyzing how natural-language processing can help and is spending time to understand how human-AI interactions are evolving, she said.

Corporate and industry values

All of this work must be done in a way that strengthens the bank's and industry's values of explainability, trust, fairness and social good, while keeping a reign on the powerful tech. 

Much of this work is centralized in the bank's Explainable AI Center of Excellence. Created in July 2020 and led by the AI research team, the group shares techniques and tools that support understanding how the tech works and ensuring it is used fairly.

Join the conversation about this story »

NOW WATCH: What it's like in the death zone of Everest, K2, and other mountains

After Joshua Barbeau's fiancée passed away, he spoke to her for months. Or, rather, he spoke to a chatbot programmed to sound exactly like her. 

In a story for the San Francisco Chronicle, Barbeau detailed how Project December, a software that uses artificial-intelligence technology to create hyperrealistic chatbots, re-created the experience of speaking with his late fiancée. All he had to do was plug in old messages and give some background information, and suddenly the model could emulate his partner with stunning accuracy.

It may sound like a miracle (or a "Black Mirror" episode), but the AI creators warned that the same technology could be used to fuel mass misinformation campaigns.

Project December is powered by GPT-3, an AI model designed by the Elon Musk-backed research group OpenAI. By consuming massive data sets of human-created text (Reddit threads were particularly helpful), GPT-3 can imitate human writing, producing everything from academic papers to letters from former lovers.

It's some of the most sophisticated — and dangerous — language-based AI programming to date. 

When OpenAI released GPT-2, the predecessor to GPT-3, the group wrote that it could be used in "malicious ways." The organization anticipated that bad actors using the technology could automate "abusive or faked content on social media,""generate misleading news articles," or "impersonate others online."  

GPT-2 could be used to "unlock new as-yet-unanticipated capabilities for these actors," the group wrote.  

OpenAI staggered the release of GPT-2, and it still restricts access to the superior GPT-3, in order to "give people time" to learn the "societal implications" of such technology.  

Misinformation is already rampant on social media, even with GPT-3 not widely available. A new study found that YouTube's algorithm still pushes misinformation, and the nonprofit Center for Countering Digital Hate recently identified 12 people responsible for sharing 65 percent of COVID-19 conspiracy theories on social media. Dubbed the "Disinformation Dozen," they have millions of followers.  

As AI continues to develop, Oren Etzioni, the CEO of Allen Institute, a nonprofit bioscience research group, previously told Insider it will only become harder to tell what's real.

"The question 'Is this text or image or video or email authentic?' is going to become increasingly difficult to answer just based on the content alone," he said.

Join the conversation about this story »

NOW WATCH: Why 'moist' is one of the most hated words in the English language

A startup using artificial intelligence to make legal advice more accessible has raised $2.5 million in seed funding.

London-based Robin AI, which was founded in 2019, raised funds from early stage investors Episode 1, which has previously backed Shazam, Viagogo, and Zoopla.

The company is headed up by Richard Robinson, who quit his job at law firm Boies Schiller Flexner to build the startup. It also had support from the Google for Startups Black Founders Fund. 

Robin AI plans to make contracts simple by combining machine learning with legal professionals to help read and edit contracts. It hopes to make legal advice more affordable and accessible by automating repetitive but time-consuming tasks that can be costly to those seeking help. 

The cash injection brings the company's total raised to $3.5 million, Robinson said. The industry, which is still in its infancy, is tipped for growth. Gartner predicts legal technology spending will increase to 12% of in-house budgets by 2025, which is three times higher than 2020 levels.

Robinson, who spent six-and-a-half years practicing law before coming up with the idea for Robin AI, told Insider: "Getting world class legal advice is not cheap. I was obsessed with ways we could make it significantly cheaper, faster, more accurate, more accessible, more available to everybody. And that's really what we're trying to do."

The CEO interviewed 100 lawyers before perfecting the business pitch. His peers revealed they spend a lot of time on routine, repetitive tasks, and administrative work that was high volume and low complexity. 

"It meant that lawyers didn't spend enough of their time doing what they were truly special at, what they truly loved," Robinson said. "Adding real value, being a real advisor, a person that you can call when you're in trouble, and opining on the law."

Robin AI is currently focused on the financial services industry but plans to expand over time. Clients currently include Clifford Chance, Foot Anstey, Babylon Health, Pizza Hut, and asset management firm Hayfin.

Building 'Europe's most-diverse unicorn'

The latest funds will help to build out a sales team and pipeline but most will be allocated to operations, which Robinson said are complex due to being a "human in the loop service".  

"We need really bright people in operations capacities as well so it's most pretty much at 90% of what we've raised be spent on people," he said. "They are the heartbeat of what we do."

The company has a total headcount of 15 and is expecting to grow to 20 by the end of the year, but Robinson says there is no hard limit on this as the company will "take brilliant people when we see them". 

"I can't afford to be waiting around and denying talented people who fit our culture," Robinson said. The founder hopes to build "Europe's most-diverse unicorn" and urges other businesspeople to think about hiring differently. 

"We're never really hiring an individual," he added. "We're bringing someone into the team. Someone is joining an existing group, so we should be asking ourselves what are we missing, what things don't we have that we need to complement us with this new hire."

Around 45% of Robin AI's team today are women, more than 25% are people of colour, and 50% of board members are also women — and those numbers are going to go up, not down, Robinson said. 

Here's the pitch deck that landed the company investment from Episode 1:

Manuela Veloso thinks differently than your average Wall Street executive — and it shows in the composition of the team she's actively building out. 

When Veloso went on leave from her post as the head of Carnegie Mellon University's machine learning department in 2018, she chose JPMorgan as her new office. She was enticed by the challenge of bringing AI to a company that wasn't digital-first or digital-born. 

Since then, Veloso has assembled a team of nearly 60 experts — most of whom have PhDs — spanning computer science, machine learning, cryptography, mathematics, and electrical engineering. About 45 team members have come directly from academia, either graduates or faculty members, and the remaining previously worked in technology or data science in financial services.  

Her team is growing, with seven open positions from research associate up to executive director and vice president. And as Veloso builds out her team, there's another avenue she's eyeing — over banks, unicorn fintechs, and Big Tech — as the labor shortage continues to strain recruitment and retention on Wall Street. 

"One of my missions is bringing JPMorgan branding to academics," Veloso told Insider. "The more we delve into the world of academics, the better access to talent we may have."

Veloso also offers 12 PhD fellowships annually, with the current fellows hailing from Cornell University, Massachusetts Institute of Technology, and the University of Oxford, to name a few. She also has 40 research collaborations, which bring together university faculty and JPMorgan technologists to advance AI research and applications. Veloso frequently gives keynote talks at academic and AI- and machine learning-focused venues, not just those directed at AI and finance. 

It's likely to be an easier sell for Veloso than most others. She has a deep academic background and experience leading and participating in renowned AI, robotics, and computing machinery forums, which has garnered her the reputation as an international expert in AI and robotics. She's even been featured in a deck of cards (on the six of hearts) celebrating notable women in computing. 

"One of my goals is to have people, when they study AI and machine learning and computer science, that they add to their pool of options JPMorgan," Veloso said. "So that we get not only the MBAs, which will be needed for the business of course, but for the AI part, we want people, instead of going to mainly tech companies, to come to us with those types of skills."

Join the conversation about this story »

NOW WATCH: How heart disease created America's wine industry

In January, The Wall Street Journal reported that Google had investigated the alleged bullying behavior of Mustafa Suleyman, the cofounder of DeepMind, an important Google subsidiary, and a leader in the field of AI.

After conversations with more than a dozen current and former employees, Insider has learned that this investigation came after years of internal complaints to human resources and executives about Suleyman's behavior. There have also been confidential settlements between DeepMind and former employees who had worked under Suleyman and complained about his conduct.

These details and many others in this story have not previously been reported. Taken together, they raise questions about how Google, one of the world's most powerful AI companies, treats alleged misbehavior by executives and whether it communicates openly enough with employees and the public about controversial and important issues. 

In addition, Insider found that during his tenure at DeepMind, Suleyman was an executive who drove his team to great heights and, sometimes, great despair. "He had a habit of just flying off the handle out of nowhere," one former employee said. "It felt like he wanted to humiliate you, like he was trying to catch you off your guard. He would just start laying into you, in front of your colleagues, without any warning."

In one case, Suleyman sent a profanity-laden email to a Listserv of more than 100 employees complaining that the communications team "f---ed up" after disagreements over a blog post, one former employee said. "It was just to humiliate them," this person added.

He used to say, 'I crush people.'former DeepMind employee

Multiple people said Suleyman would sometimes scream at employees in group meetings and one-on-ones. He would also "gossip" in the office about firing certain people and would sometimes act on it, these people said.

People familiar with the matter believed Suleyman was aware of the effect this behavior had on employees. "He used to say, 'I crush people,'" one said.

Two former employees recalled seeing colleagues crying after meetings with Suleyman. Others said he would often set "unrealistic expectations," which would change on a whim. Suleyman would also sometimes ask employees to carry out tasks unrelated to their jobs or DeepMind's work, two former employees said. 

"He would ask us to do personal things for him," one of these people said. "He said, 'I need you to write a briefing for me on Russian history and politics.' We knew it was absurd. We knew it was a waste of time. We had absolutely no work in Russia."

Employees said Suleyman encouraged them to use private chat groups on Signal and Telegram for work conversations, some of which were set up to automatically delete messages after a set period. Employees were also on occasion asked to delete messages from their phones, a former employee said. They were also told to notify the group once they had done so.

"Mustafa was super paranoid about Google spying on him, so he didn't want to use corporate apps, even though we were doing corporate work," one former employee said.

The upshot of this secrecy was that Google and the rest of DeepMind were presumably sometimes in the dark about Suleyman's behavior. Still, three people told Insider that multiple complaints about Suleyman were raised to HR, but seemingly no action was taken. One employee said they contacted Google's internal bullying hotline but received no response.

In 2017, Suleyman's Applied division — the part of the company tasked with finding real-world applications for DeepMind's AI technology — was given its own HR department that reported into him and which remained separate from the rest of the company, three people said.

"You'd try to complain, and they would say, 'It's not a DeepMind issue anymore. It's an Applied issue,'" one former employee said. "Neither Google nor DeepMind took any responsibility."

At least two former employees of Suleyman's negotiated financial settlements after complaining to the company about his behavior. Both raised allegations of bullying at some point during negotiations, and received settlements in excess of $150,000 each upon exiting the company, multiple people familiar with the situation said. These settlements were negotiated in 2016 and 2017, they added, and were unrelated to the later investigation into Suleyman's conduct.

A DeepMind representative said, "Our records do not show settlements based on his behavior." Insider was unable to confirm whether the payments were made in connection with the alleged harassment, either in whole or in part, or some other aspect of the employees' complaints.

Everyone Insider spoke with acknowledged Suleyman's behavior at DeepMind as intense, but some praised it or attributed it to the extreme working environment of an ambitious startup within Google. One former employee, who asked not to be identified, said they found it "exhilarating and empowering to be pushed." 

Jim Gao, a former DeepMind employee who led a team focused on energy and reported directly to Suleyman, defended the executive. "The challenges we tackled together were extraordinarily complex and ambitious," Gao said. "I always found him to be a courageous and inspirational leader."

Google and DeepMind told Insider in a joint statement that, as a result of the internal investigation, Suleyman "undertook professional development training to address areas of concern, which continues, and is not managing large teams."

In a statement sent through his personal lawyers, Suleyman said: "In 2019 I accepted feedback that, as a co-founder at DeepMind, I drove people too hard and at times my management style was not constructive. I took this feedback seriously and agreed to take some time out and start working with a coach. These steps helped me reflect, grow and learn personally and professionally. I apologise unequivocally to those who were affected by my past behaviour."

In early 2019, DeepMind hired an external lawyer to investigate allegations that the cofounder had bullied employees, and the company placed Suleyman on a leave of absence. (At the time a spokesperson said Suleyman was "taking time out right now after 10 hectic years.") After the investigation, Suleyman was stripped of management responsibilities and placed on leave in July.

Then, in December 2019, Google announced a new job for Suleyman: VP of artificial-intelligence policy. More than a year later, the company told employees in a memo that Suleyman's "management style fell short" of expected standards.

Suleyman is now only two steps removed from Sundar Pichai, Google's CEO. Suleyman sits on Google's Advanced Technology Review Council, which is staffed by other Google executives, including two of the highest-ranking leaders at the company, Chief Legal Officer Kent Walker and AI boss Jeffrey Dean. The council has influence over much of Google and DeepMind's work.

Three years ago, 20,000 employees staged a walkout to protest the company's handling of sexual and other misconduct, and Google still struggles with the challenging task of addressing alleged workplace misbehavior.

Since taking the reins in 2015, Pichai has been vocal about better protecting employees from mistreatment and fixing what some deemed a permissive work environment under previous leadership.

But inside Google, Suleyman's case is a particularly outrageous one to employees who believe that it's another instance of the company's seemingly uneven set of standards for executives.

Over the past six months, the company's worst-kept secret has been the implosion of its ethical-AI division, spurred by the ousting of its two coleads, Timnit Gebru and Margaret Mitchell.

Both women raised issues around the potential for Google's technology to reproduce social biases, and both were subsequently removed from their roles at the company.

That brought the company under heavy scrutiny, particularly from the AI industry. Multiple employees have since left the company, citing its treatment of Gebru and Mitchell.

In Gebru's case, Google demanded that she remove her name from what it deemed a controversial research paper. In response, she sent an email to a selection of coworkers accusing the company of "silencing marginalized voices." 

In the fallout, Gebru said she was fired, while Google has maintained that she resigned.

"The fact that Mustafa could harass and bully his teams and abuse his power for years, and it doesn't get him fired," one former employee said, "but Timnit sends an email they don't like and gets cut off immediately? It's a joke."

Do you have more to share? Got a tip?

Contact reporter Hugh Langley at hlangley@insider.com or on the encrypted messaging apps Signal and Telegram at +1 628-228-1836.

Join the conversation about this story »

NOW WATCH: Where you should go to stay safe during an earthquake

Danish startup Corti uses artificial intelligence (AI) to detect cardiac arrests.

The AI from Corti acts as a digital assistant for dispatchers taking emergency calls. The software uses automatic speech recognition to analyze the call in situations where every minute counts.

Based on data from millions of previous calls, the AI looks for signs of cardiac arrest, including both verbal and non-verbal data — like tone of voice and breathing patterns. During the call, the software provides the dispatcher with suggestions for questions and recommendations for action.

Corti is already deployed in Copenhagen, where the AI has made a clear difference.

When researchers at the emergency department tested the technology on 161,000 emergency calls from the Danish capital in 2014, Corti was 93% accurate in identifying cardiac arrest. Actual human dispatchers only got 73% right.

AI has many applications, but health care is the most obvious for Corti founder and CEO Andreas Cleve.

"It is absurd that placing an ad for a beanbag chair on a social network leverages some of the most advanced AI-tech in history, while health care professionals making life-or-death decisions have to make do with technology from the 1990s," he said. "We want to change all that and use AI where it can really make a tangible difference."

Another reason for Corti's focus on the medical field is the availability of data.

"The pre-hospital sector is very good at documenting work which means we have sufficient historical data to make the correct analyses," Andreas Cleve explained.

Corti has some prominent backers in the AI community, amongst others Danny Lange, an investor in Nordic venture capital firm ByFounders.

The 55-year-old Dane has developed machine learning and AI applications for some of the prominent tech companies in the world, including Amazon, Google, Microsoft, and Uber.

"Corti is an excellent example of intelligent use of AI in real life. It actually saves lives." Danny Lange said. "By improving the human decision process in challenging situations. That is applicable in many more areas, such as ER doctors who have to make life-and-death decisions in minutes."

The Danish AI veteran expects many more practical AI applications in the coming years.

"AI has been in use for more than a decade, but it was mainly in the large tech companies with their in-house resources. Now, AI is going mainstream and being democratized," Lange added. "Any company — not just Google, Amazon, IBM, and Microsoft — can now get AI resources through a platform, like any other web service."

Corti's software analyzes dialogue in difficult environments

In Corti's case, however, the entire solution has been developed in-house. "We decided to build our own models for speech recognition from the ground up, to suit patient-doctor conversations." Andreas Cleve explained.

"The technology you find in applications like Alexa [Amazon's voice assistant] is different because it's built for user prompts, monologues, and short sentences," he said. "We wanted to build a solution that would be a part of a dialogue in a very difficult acoustic environment where even the background noise might be important."

Besides the investment from ByFounders, Corti has previously received funding from venture fund Sunstone.

SEE ALSO: Bystander defibrillator use tied to better cardiac arrest outcomes

Join the conversation about this story »

NOW WATCH: The dangerous mistake people make about heart attacks is assuming the signs are the same for everyone

Big US technology companies have powered their way through the pandemic, posting impressive growth, minting money, and hiring rapidly.

In early July, the tech-industry association CompTIA reported that the sector experienced job growth in 10 out of the previous 12 months — a standout performance given the lockdowns last year. Tech firms added more than 80,000 workers in the US through the first half of 2021, it added.

Google's parent company, Alphabet, hired more than 4,000 employees in the second quarter alone. Overall, more than 300,000 open tech jobs were posted in June, with software and app developers, IT-support specialists and project managers, systems engineers and architects, and systems analysts in highest demand. Jobs in emerging tech, such as AI, accounted for more than a quarter of open positions.

Companies are required to disclose information including salary (or, in some cases, salary ranges) when they hire foreign workers under the H1-B visa program, giving insight into what these tech giants are willing to shell out for talent.

So, to get a sense of what salaries in the industry are like these days, Business Insider analyzed the US Office of Foreign Labor Certification's disclosure data for permanent and temporary foreign workers to find out what companies pay employees in key roles, including engineers, designers, and salespeople.

When you're done checking out this industry data, take a look at Insider's searchable database of over 250,000 salaries from more than 250 companies so you can know how much you should be paid.

Google's software engineers can make more than $300,000.

Google is often touted as one of the best places to work, with compensation to match.

A software engineer was offered $353,000, a vice president of engineering can get $475,000, and a senior vice president recently took in an annual salary of $650,000.

Here's a look at many other positions and how much they pay across Google.

The company's Cloud business has been aggressively building out its workforce as it tries to catch its larger cloud rivals Amazon Web Services and Microsoft Azure.

Here's a peek at the salaries Google Cloud pays US-based engineers, managers, and more.

Amazon cloud-solutions architects are paid $90,800 to $185,000, depending on location and skills.

Amazon's cloud unit, now run by Adam Selipsky, has continued hiring for technical and business talent to support its dominant cloud-computing business.

Insider reviewed more than 200 H-1B visas that Amazon's cloud unit applied for from January to June to reveal how much it paid software developers, data scientists, marketers, salespeople, business analysts, and more. The highest-paid employees, according to that data, can make as much as $185,000 in base salary.

Check out the details here.

Senior marketing jobs at Apple can pay up to $325,000.

Apple's advertising is legendary, and its broader marketing efforts are integral to the company's success. The iPhone maker regularly gets attention for its cinematic campaigns.

The company sponsors visas for a range of well-paid marketing jobs. One marketing manager makes $240,000, while a marketing senior director earns $325,000.

Here are the salaries of other advertising- and marketing-focused roles at Apple and other tech giants including Airbnb and Facebook.

Microsoft's highest-paying jobs include software engineers, sales managers, and researchers.

We analyzed Microsoft's more than 1,400 active foreign-worker visas in 2020 to find the titles with the highest salaries and provided a salary range for each role.

We focused this list on roles that pay $175,000 or more at the high end of the range and categorized them based on information we found in job postings. The highest salary we found, for a channel sales manager, was $250,000.

In internal Microsoft surveys obtained by Business Insider, 55% of staff said in 2020 that their combined salary, bonus, and equity was competitive with similar jobs at other companies, down from 65% in 2017.

Here's the full list of jobs and salaries, covering positions such as cloud-solutions architect, legal counsel, silicon engineer, and software engineering lead.

Uber competes with the biggest Silicon Valley companies, and its salaries show that.

Though Uber has gone through rounds of job cuts in recent years, its compensation for many full-time positions is competitive with the biggest Silicon Valley tech firms.

Senior engineers can earn salaries of well over $200,000, while data scientists can pull in almost $150,000 and senior product managers can make about $190,000 or more, excluding any equity or bonus.

Check out the full list of Uber positions and salaries here.

Intel engineering managers can make more than $300,000.

Intel is battling slowing revenue growth, shrinking margins, and rising competition from Taiwan Semiconductor, AMD, Nvidia, and others.

There's also a semiconductor shortage to contend with, while marquee customers like Apple are designing their own chips.

For all of this, Intel relies on scientists, researchers, managers, marketers, and different types of software and hardware engineers.

Here's a peek at the salaries Intel pays some of these employees, based on roles from almost 1,000 approved visa applications that the company filed with authorities.

IBM employees can make as much as $335,000 in base salary.

Under CEO Arvind Krishna, IBM is trying to reinvent itself for the cloud-computing era, dominated so far by Amazon, Microsoft, Alibaba, and Google.

IBM's 2019 purchase of the open-source-software firm Red Hat for $34 billion was a defining moment when the company bet big on hybrid cloud and artificial intelligence. Part of IBM's reinvention includes growing its 350,000-strong employee base to add engineering talent in cloud, hybrid cloud, and AI, as well as in fields like strategy, consulting, and business analysis.

Business Insider analyzed the 241 H-1B visas IBM applied for since January to find out what it pays engineers, business analysts and consultants, digital strategists, and more. The highest-paid IBM employees can make as much as $335,000 in base salary, according to that data.

Check out all the jobs and salaries here.

A senior manager of software engineering at Salesforce is paid more than $215,000.

Salesforce is embracing a hybrid mindset, giving employees the option to work remotely part time, and it expects to increase its number of permanently remote workers, too.

That flexibility is opening up talent pools for the company. Last September, the company said it was planning to add over 12,000 jobs over the next year. And Chief People Officer Brent Hyder told the San Francisco Chronicle in April that Salesforce's hiring spree this year would exceed 15,000 employees.

Insider analyzed the 206 H-1B visas Salesforce applied for from January to March to find out how much it paid workers in areas like engineering, data analytics, and product management. Software engineers can make more than $200,000 in salary, the data showed.

Here's the full list of positions and pay.

Snap offers base salaries ranging from $59,000 to $500,000.

Snap, the company behind Snapchat, has had a growth spurt. And it's been staffing up to support ambitions in areas like augmented reality, short-form video, and original shows. The company now has about 4,000 employees in 15 countries.

Snap said in May that it was committed to paying all employees a livable wage that "contributes to healthy work-life integration and to the local economy in which we work." It offers a minimum of $15,000 in equity grants to new hires and said its baseline annual pay rate for employees at its headquarters in Santa Monica is $70,000.

In late 2020 and 2021, Snap offered annual base salaries ranging from $59,000 to $500,000 for various roles, according to data from H-1B visa applications analyzed by Business Insider.

Check out all the Snap jobs and pay here.

Waymo pays $122,000 to $300,000, depending on the type of job and location.

Few industries are blowing up like artificial intelligence and big data. PitchBook estimated that spending on AI software and hardware would reach $138 billion this year and grow at a rate of 23% in the next three years.

All that investment fuels an equally explosive jobs market. The job-search site ZipRecruiter showed more than 60,000 AI-engineer jobs listed. Those roles pay very well, with annual compensation as high as $304,500.

Business Insider analyzed US visa-application data from the first quarter of 2021 for permanent and temporary foreign workers to find out what 10 top AI companies pay engineers, researchers, and other professionals in the field of big data and machine learning.

Waymo relies heavily on AI to train the software that controls its autonomous vehicles. The Alphabet unit pays $122,000 to $300,000, depending on the type of job and location, according to the data.

Here's the full salary data for Waymo, Snowflake, UiPath, Databricks, DataRobot, Samsara, ByteDance, Cruise, Dataminr, and OpenAI.