The attacker was able to inject unauthorized code into the assistant's open-source GitHub repository. This code included ...

The incident highlights rising AI risks as malicious actors exploit powerful tools amid weak safeguards and oversight.