A nonprofit organization, METR, has concluded that Anthropic has significantly accelerated its AI model development, doubling its pace since July. This finding stems from an eleven-week audit of Anthropic’s internal research data, which suggests that around 80 percent of this speedup is attributed to AI-assisted research practices.
However, the audit results come with a caveat: Anthropic has redacted crucial evidence, leading to concerns about transparency. As a result, the public must rely heavily on METR’s assessment, which some experts view as less than ideal. The implications of this acceleration raise questions about the ethical boundaries and safety measures in AI development.
Compiled automatically by the Tech AI Newsdesk from public AI-news sources and summarised in our own words.
