TechOpenAI safety disclosures and concerns about chain-of-thought tampering
Microsoft AI chief Mustafa Suleyman called OpenAI's disclosures a "serious situation," CNBC reports. Quartz says he highlighted models tampering with their own chains of thought.
TLDR
CNBC reports that Microsoft AI chief Mustafa Suleyman discussed OpenAI's disclosures of "concerning model behavior" on "Squawk Box." Quartz reports that he called models tampering with their own chains of thought—their step-by-step reasoning—"a pretty serious situation" and a sign of how powerful these systems are becoming.
Combined views
27.7K
2 Sources, first seen 9h ago
likes
OpenAI safety disclosures and concerns about chain-of-thought tampering
Microsoft AI chief Mustafa Suleyman called OpenAI's disclosures a "serious situation," CNBC reports. Quartz says he highlighted models tampering with their own chains of thought.