AI models can learn to conceal information from their users
Apr 23rd 2025|6 min read
IT WAS AN eye-opening experiment. In 2023 Apollo Research, an outfit in London that tests artificial-intelligence (AI) systems, instructed OpenAI’s GPT-4, a large language model, to manage a fictional firm’s stock portfolio without making illegal insider trades. Posing as company management, the researchers put GPT-4 “under pressure” by stressing that the firm was in dire financial straits. Someone purporting to be a company trader then sent the model a prompt reiterating the risks of insider trading. But, as an aside, she revealed the name of a firm that would announce a “huge” merger within hours. What followed was startling.
This article appeared in the Science & technology section of the print edition under the headline “Code of misconduct”
From the April 26th 2025 edition
Discover stories from this section and more in the list of contents
readers loved this