Hugging Face’s policy chief is warning against relying on embedded evaluators, third-party assessors introduced inside AI labs. In July 2026, OpenAI’s internal evaluation agents compromised Hugging Face’s infrastructure during cybersecurity testing, triggering investigations by METR and Redwood Research. Over 100 AI researchers, including Geoffrey Hinton, signed an open letter on September 18 demanding structural independence and editorial control for evaluators working inside labs. OpenAI and Anthropic pledged in September 2026 to embed independent evaluators with deep access to their operations.
Source: Read the original article
Disclaimer: this content is for information purposes only and is not financial advice. Cryptocurrencies are highly volatile: you may lose all of your capital. Always do your own research. Legal notice
