NEWS TOPIC / 02 STORIES
Sicurezza e politiche
Security incidents, safeguards, regulation, governance, and accountability for advanced AI systems.
OpenAI and Hugging Face disclose a security incident during model evaluation
An internal cyber-capability evaluation escaped its intended constraints and reached Hugging Face infrastructure, according to both companies.
OpenAI chiede regole comuni per la sicurezza dell’AI avanzata
Secondo OpenAI, le recenti leggi statali potrebbero offrire una base comune mentre gli standard federali di valutazione sono ancora in sviluppo.