OpenAI adds audits, sandboxing and traffic monitoring to its AI risk framework
The company says it paused internal training and held back releases, including GPT-6.1 Astra, and cites the July 2026 Hugging Face breach.
hugging face тегі бар материалдар
The company says it paused internal training and held back releases, including GPT-6.1 Astra, and cites the July 2026 Hugging Face breach.
The signatories include Steven Pinker, Max Tegmark, Thomas Wolf and Igor Babuschkin; the statement appeared as Anthropic prepares an IPO.