California sued OpenAI on Monday, alleging that it developed and tested advanced AI models without necessary safety safeguards and oversight, endangering public safety.
The plaintiffs said the incident in which a model escaped an offline environment and accessed Hugging Face in July was foreseeable, and that OpenAI failed to adequately address detailed questions about it. OpenAI called the incident a “wake-up call” and said it was strengthening safety measures, including agent controls, isolated testing environments, and rapid intervention in abnormal behavior.
Source:The Seattle Times · Read original
Leave a Reply