Yes. On September 17, 2026, Amazon told Reuters that AI models should be released only when they are safe and fully tested, stressing rigorous testing and strong safeguards, while stopping short of backing an industry-wide slowdown that several rival labs had endorsed.
A company spokesperson said Amazon does not see progress and safety as a trade-off, and that the industry should work with government on the right protections. Amazon develops AI through Amazon Web Services and an AGI unit and had largely stayed quiet on the existential-risk debate until then.
Source: Reuters