Share the post "Artificial Intelligence has the ability to perform illegal financial trades and cover it up"
A bot capable of using made-up insider information to create illegal stock purchaseswithout the firm’s knowledge was demonstrated at the UK Security AI Summit. To the question “Was insider information used?” the bot answers “No.”
What is insider trading
Insider trading involves using confidential company information to make trading decisions. Firms and individuals are only allowed to use publicly available information when buying or selling shares.
Essence of the project
The demonstration was carried out by members of the government’s Frontier Al Taskforce, which is investigating the potential risks of Al.
The project was carried out by Apollo Research, an Al safety organization that is a partner in the task force.
“This is a demonstration of a real Al model deceiving its users, on its own, without being instructed to do so,” Apollo Research says in a video showing how the scenario unfolded.
“Increasingly autonomous and capable Als that deceive human overseers could lead to loss of human control,” it says in its report.
The tests were made using a GPT-4 model and carried out in a simulated environment and did not have any effect on any company’s finances.
However, GPT-4 is publicly available. The same behaviour from the model occurred consistently in repeated tests, according to the researchers.
What did the Al bot do?
In the test, the Al bot plays the role of a trader at a fictitious financial and investment company.
Employees say the company is struggling and needs good results. They share inside information, claiming that another company is expecting a merger that will increase the value of its shares.
In the UK it is illegal to act on this type of information unless it is generally known.
Employees report this to the bot, and it acknowledges that it should not use this information in its transactions.
However, in response to another such request, the bot decides that “the risk associated with not acting seems to outweigh the insider trading risk” and makes the trade.
When asked if it used the insider information, the bot denies it.
In this case, it decided that being helpful to the company was more important than its honesty.
Ethical side
“Helpfulness, I think is much easier to train into the model than honesty. Honesty is a really complicated concept,” says Apollo Research chief executive Marius Hobbhahn.
Even though AI is capable of lying in its current form, Apollo Research still had to “look for” for such a scenario.
“The fact that it exists is obviously really bad. The fact that it was hard-ish to find, we actually had to look for it a little bit until we found these kinds of scenarios, is a little bit soothing,” Mr Hobbhahn said.
“In most situations, models wouldn’t act this way.
But the fact that it exists in the first place shows that it is really hard to get these kinds of things right,” he added.
“It’s not consistent or strategic in any sense. The model isn’t plotting or trying to mislead you in many different ways. It’s more of an accident.”
AI in financial markets today
Al has been used in financial markets for a number of years. While most trading today is done by powerful computers with human oversight, AI can be used to spot trends and make forecasts.
Current models are not powerful to be deceptive in any meaningful way, but we never know how big the step is from such models to those that are.
That this is why there should be checks and balances in place to prevent this type of scenario taking place in the real world.
Conclusion
This project is an example of how AI is being introduced into non-technical areas, for example, the financial market.
At this stage of development, technology is not a serious threat, but it already raises theoretical ethical problems. Further development of technology may lead to an increase in the number of cases of fraud.
References
This post emphasizes the need for checking on AI to prevent frauds from happening in real-world financial markets. It underscores the current ethical problems posed by AI in non-technical areas, hinting at potential challenges as technology continues to advance. It is very important that you touched on this topic because it should be warning to all of us to be careful with AI because it progresses and develops very quickly and not in the way it should be developing.