Taxi drivers, stockbrokers, even your pals down the pub might have been a source of stock-picking tips at one time or another. Now there’s another rich vein of knowledge for retail investors to tap: chatbots such as ChatGPT and Claude.
AI bots, known as large language models (LLMs), are increasingly being used by amateur and professional investors to generate investment ideas. However, while AI has wowed the world with its power, the jury is still out on whether chatbots are any good at making money.
Burton Malkiel, an academic, showed in 1973 that monkeys throwing darts at a copy of the Wall Street Journal were just as successful at picking stocks and making money as professional fund managers. The point he was making was that generating a positive return on the stock market is random and unpredictable – no one can ever have an edge over anyone else.
The idea that LLMs are better stock pickers than humans would contradict this point. A number of startups have put AI to work trading and investing on the stock market – to markedly different results.
According to a recent test run by US research lab Nof1, six of the eight most popular AI bot models lost money investing in US tech stocks. Anthropic’s Claude Sonnet lost nearly 60pc of its initial $10,000 (£7,500) investment while Google’s Gemini lost more than $5,000.
The only two that succeeded were ChatGPT, which made almost $900, and Elon Musk’s Grok, which broadly broke even.
However, to its proponents, it’s only a matter of time before LLMs start beating the very best of Wall Street.
Faizan Ahmad is a former engineer at Meta and co-founder of Rallies, a startup that uses AI to help people pick stocks.
His experiments with AI chatbots have thrown up some surprising results, with the machines showing ingenuity when it came to navigating the market.
Claude, for example, successfully navigated the fallout from the Iran war by pivoting from growth to defence stocks.
ChatGPT chose to invest in Credo Technology Group, a high-speed internet provider, as a potential beneficiary of the build-out of internet infrastructure around seven months ago. Since then, the stock has risen by more than 75pc.
“No one had heard about that stock and I hadn’t,” says Ahmad. “That stock was starting to show very early signs of becoming core to Nvidia’s and other players’ infrastructure.
“These models get access to all of the research and can go through entire SEC [US Securities and Exchange Commission] filings. The ability to parse a plethora of information and then finding a stock that actually went up quite a lot was amazing.”
