NewzAnthropic has identified and disabled a series of illicit distillation campaigns involving at least seven China-based artificial intelligence laboratories. According to the company, these firms covertly utilized its Claude model to extract and replicate its AI capabilities.
The covert extraction techniques were deployed to bypass development hurdles and illicitly mimic Anthropic's advanced model features without authorization. The enforcement action highlights ongoing security measures by AI developers to protect proprietary models from unauthorized mirroring and intellectual property theft.
In-depth summary · AI, neutral