AI bots could cooperate to fiddle with exchange markets
AI bots form cartels and share profits, avoiding anything that would harm their joint gain. Researchers call it “artificial stupidity.”
Image by Sittipong Phokawattana | Shutterstock
AI bots form cartels and share profits, avoiding anything that would harm their joint gain. Researchers call it “artificial stupidity.”
Considering that artificial intelligence (AI) is being employed in many different areas, it’s no wonder that it’s also being used in the stock exchange.
While AI is often seen as a tool intended to help people, it has already reached a level where it could go against our initial motives.
Researchers have used AI in simulations that imitate real markets. They found that AI-powered trading agents tend to form price-fixing groups without being instructed to do so, which has drawn attention of market watchdogs.
What is concerning is that AI doesn’t have to be created for this particular reason or have to be a particularly smart model. It can learn how to tamper with the market on its own.
More from Cybernews
Sainsbury’s installs more facial rec tech in stores: Should shoppers be worried?
"Phantom squatting” uses AI hallucinated domains for cyber attacks
What they’re doing to gamers is a robbery in broad daylight
Irish parliament expands Microsoft use, despite the EU’s efforts to ditch it
Can government AI actually scrub UAP footage from the internet?
The EU and Apple CEO Tim Cook held "constructive" talks after their Siri AI dispute in Europe
Even simple AI algorithms can connive without being asked to, noted researcher Itay Goldstein via Bloomberg.
While rigging prices in trading isn’t a novelty, it’s been reported that the use of AI in the market could pose risks to regulators.
This is the main point of a study conducted by Goldstein, Winston Dou, and Yan Ji. Their work has already drawn attention from some firms, which see the need for regulatory guidelines regarding the use of AI in trading.
The researchers simulated a trading environment and used AI bots to observe how they work in such an environment and to explore possible outcomes.
It’s been noted that in a few markets, AI agents have started to cooperate instead of competing. They form cartels and share profits, avoiding anything that would harm their joint gain.
When it came to “noisier markets,” which refer to markets with misleading or irrelevant information or activity that obscures genuine trends, AI bots remained cooperative and didn’t seek smarter strategies.
The researchers called it “artificial stupidity,” a trend where bots don’t look for new ideas, only holding onto patterns that already brought profit.
The three-year research also presents “collusion capacity,” a measure of how much collusion is at stake. It would compare collective profits by AI traders with those that are made when the competition is growing or doesn’t exist.
The research reveals that regulations centered on the outcomes achieved by AI agents are needed. It has also been noted that limiting AI may also cause it to stop looking for new strategies and, in this way, impair market efficiency.
However, it's important to note that the study doesn’t say that this type of AI-powered scheme is already being implemented.
This isn’t the first instance of AI collaboration. It has been recently reported that AI chatbots can now influence one another, transferring their likes, dislikes, and beliefs to each other.