The Threat Intelligence Group at Google found that account credentials for multiple artificial intelligence models are being traded on illegal marketplaces. These listings offer unauthorized access to models from OpenAI, Anthropic, and Google with prices slashed by as much as 97%, the Financial Times reports.

The activity is part of a broader surge in cybercrime targeting AI infrastructure, according to the Financial Times. Hackers are increasingly focusing their efforts on compromising AI servers and user accounts to acquire and resell high-value access to the tools.