According to @timnitGebru, Anthropic and Accenture will invest $1B in embedded frontier AI evaluation over five years, strengthening independent testing.

According to @GoogleDeepMind, it is piloting double-blind frontier AI testing to keep prompts and model weights private and enable robust external audits.

According to FoxNewsAI, industry calls grow for stricter AI testing safeguards and enforceable accountability after reported real‑world harms.