- OpenAI has disrupted a major coordinated attack on ChatGPT involving more than 15,000 users.
- The attack was an attempt to distil OpenAI's AI models to extract and copy the underlying technology.
- OpenAI implicated Chinese AI firm Moonshot AI in part of the activity, which escalated throughout July before being stopped on 28 July.
- The company warned that adversarial distillation poses national security and safety risks by copying models without safety safeguards.
- OpenAI mitigated the threat by terminating suspect accounts, fixing security bugs, and sharing intelligence with governments and industry peers. Moonshot AI did not immediately respond to a request for comment from The Independent.
IN FULL
