Who to sue when a robot loses your money

Original author: Thomas Beardsworth and Nishant Kumar
  • Transfer

The first case when people went to court about the loss of investment, which resulted in the work of autonomous machines, will be a check of the limits of reliability

Every day, robots are becoming more and more like humans, but they still can not sue them.

Therefore, the Hong Kong tycoon chose a different option for lack of a better one. He is suing the seller, who convinced him to entrust part of his capital to a supercomputer, the exchange of which cost him more than $ 20 million.

The case came up against Samatur Lee Kin-kan, whose father is one of the largest investors in Shaftesbury Plc, which owns most of London's Chinatown, Covent Garden and Carnaby Street, against Rafael Costa, who spent most of his career selling investment funds to such companies like Man Group Plc and GLG Partners Inc. This is the first time that people have filed a lawsuit about the loss of investment caused by the operation of autonomous machines, which will be a check of reliability limits, and that will bring to light the “black box” problem: if people don’t know how the computer makes decisions, who answers when will something go wrong?

“People think that algorithms work faster than people and make better decisions,” said Mark Lemley, a professor of law at Stanford University who runs the local Law, Science and Technology program. “Often it is, but when it turns out to be wrong, or when the algorithms go astray, investors are looking for a scapegoat.”

Rafael Costa

The deadlines preceding the lawsuit were taken from documents sent to the commercial court in London, where the hearing should begin in April next year. And it all started over dinner at a restaurant in Dubai on March 19, 2017. Then Lee, 45 years old, met with Costa, 49-year-old Italian, known in his midst as "Captain Wizard". While eating, Costa described a robotic hedge fund, whose services for fully automatic money management using AI will soon begin to be offered by his London company Tyndaris Investments.

The K1 supercomputer, developed by the Austrian company 42.cx, will comb online sources of information such as news articles and social networks to gauge investor sentiment and make predictions about the future value of futures. Then he will send stock management instructions to the broker, eventually adjusting the strategy based on the information received.

The idea of ​​a fully automatic money manager instantly inspired Lee. He met with Costa for dinner three days later, and before that he wrote to him in an email that the AI ​​Foundation was “exactly what I needed.”

In the following months, Costa shared with Lee the results of simulations in which K1 achieved double-digit returns on invested money, although now they argue about the thoroughness of historical testing. Lee eventually entrusted K1 with the management of $ 2.5 billion - of which $ 250 million was his own, and the remainder was leverage from Citigroup Inc. This amount was planned to double.

However, Li’s K1 craze immediately passed when the computer started bidding at the end of 2017. By February 2018, he was steadily losing money, including the case when $ 20 million was lost in one day - February 14 - due to stop loss [“moose” in the jargon of traders / approx. trans.], which, according to Lee’s attorneys, would not have been enforced if K1 were as sophisticated as Costa claimed.

Lee has now filed a lawsuit against Tyndaris in the amount of $ 23 million due to an exaggeration of the capabilities of the supercomputer. Lawyers at Tyndaris, who sued Lee for $ 3 of unpaid fines, deny the fact that Costa was exaggerating K1. They argue that he never guaranteed that the AI ​​strategy would be profitable.

Sarah Makatomaini, a lawyer for investment firm Lee suing Tyndaris, declined to comment. Rob White, a spokesman for Tyndaris, declined to associate the editorial board with Costa for an interview.

Legal battles are a harbinger of what can happen when AI is included in all areas of life, from robomobiles to virtual assistants. When technology misfires, the question of who to blame for this is debatable. In March, US prosecutors acquitted Uber Technologies Inc., accused of the death of a 49-year-old pedestrian who died under the wheels of one of her robomobiles.

Profit in 2019 for every $ 100 investment made in 2014; S&P 500 earnings are based on reinvested dividends.

In the world of hedge funds, attempts to use AI have become a necessity after many years of unsatisfactory work of people-managers. Quantitative investors — computers specifically designed to search for and execute trades — are already popular. Pure AI funds that automatically learn and improve efficiency based on experience that do not have specially programmed data are less common. As soon as the AI ​​develops its own understanding of the situation, even its creators cease to understand why it makes certain decisions.

“You may find yourself in a situation where you cannot explain why you hold this position,” said Anthony Todd, co-founder of Aspect Capital, a London-based company experimenting with AI strategies before allowing them to work with client investments. “One of our concerns with using machine learning techniques is the loss of clear hypotheses about market behavior.”

Lee’s lawyers claim that Costa gained credibility by inflating the skills of technicians who wrote the K1 algorithm and said, for example, that they were developing Deep Blue, an IBM chess computer that marked the dawn of the AI ​​era when he beat the world champion in 1997- m

Garry Kasparov plays against IBM Deep Blue in 1997

In a conversation with Bloomberg, 42.cx founder Daniel Matts said that none of the computer science experts who consulted him had anything to do with Deep Blue, but one of them, Vladimir Lvovich Arlazarov , developed the Kaissa chess program in the USSR in the 1960s. "[She won the 1st World Computer Championship in Stockholm in 1974 / approx. transl.]. He acknowledged that such an experience might not be particularly strongly related to investment. Algorithms have learned quite well how to beat people in games, because they have clear rules that can be simulated - which cannot be said about the stock markets. Alazarov told Bloomberg that he gave Matts general advice, but did not work on K1.

Inspired by a study by the European Central Bank in 2015, which examined investor sentiment on Twitter, 42.cx created software that could send out mood signals, said Matts, who recently agreed to pay $ 17 million to the U.S. Securities Commission to settle a fraud lawsuit investors of their mobile payment company Jumio Inc., committed in the early 2010s. Tyndaris decided how to respond to these signals, he said.

“It was a great program,” Matts told us over the phone. “The signals we gave had clear scientific justifications.” I think we worked pretty well. I know that I am able to catch the mood. I am not a trader. ”

The court documents contain a lot of controversy over whether Lee deceived with regards to the capabilities of K1. For example, the machine created a single application in the morning if it received a clear signal about the mood of investors, although Lee claims to have considered it to be engaged in trading throughout the day. Costa's lawyers say he told Lee that buying or selling futures based on several trading signals was part of the company's plans, but was not planned from the very beginning.

There were periods of several days when K1 did not trade at all, because it did not find strong enough trends. In one of the messages to Costa, he complained that K1 was resting, while calmly skipping adverse events, in the hope that he would not reach the activation of a stop loss order (a predetermined level, upon reaching which the broker would sell shares to limit losses in a sudden fall prices).

This happened on the day of St. Valentine in 2018. In the morning, K1 placed a request with a Goldman Sachs Group Inc. broker. At $ 1.5 million in S&P 500 futures, predicting an increase in the index. But he went in the opposite direction, when it turned out that inflation in the USA grew faster than expected, which led to the activation of K1 stop loss by 1.4% and a loss of $ 20.5 million by the fund. However, after a few hours S&P regained its position - and this, according to Lee’s attorneys, shows that the stop loss border for K1 was set to “rude and wrong.”

Lee states that he was assured that K1 will use its “deep learning opportunities” daily to determine the correct stop loss margin based on market factors such as volatility. Kosta denies saying such a thing and claims to have told Lee that people will always appoint this level.

In an interview, Matts said that K1 did not lay the possibility of setting a stop loss independently - he only knows how to give out two types of mood signals: generalized, which Tyndaris can use to enter a position, and dynamic, which can be used to exit or change a position. Although Tyndaris touted a fund managed by K1 to other investors, the spokesman declined to answer the question of whether he managed any real funds. And all references to the supercomputer disappeared from his site last month.

Investors, and in particular Marcus Storr, say they are worried about the entry of AI funds into the stock exchange, especially considering that funds that use AI as their main strategy earned less than half of the S&P 500 profit over three years, according to the indices by Eurekahedge AI Hedge Fund.

“We can't judge the code,” said Storr, the investment decision maker at Bad Homburg, a German trust fund. “So we can only judge the research capabilities and structure.”

But what happens when companies use chatbots to sell products to customers? It may not even be possible to sue the seller, added Karishma Parokha, a London lawyer for Kennedys, specializing in product liability. “Misunderstanding refers to what a person tells you,” she said. “What happens when someone sells something to you?”

Also popular now: