Weekly AI News - Oct. 3, 2025

This week on our show, the hosts cover major model releases, including Anthropic's Claude Sonnet 4.5 and OpenAI's Sora 2. They then discuss new policy and industry adoption, as California passes the first comprehensive AI safety law, and Reuters launches AI-generated earnings summaries. Shifting to the enterprise, OpenAI has introduced a new metric for AI's value in the workplace while Microsoft has added powerful agent features to its 365 Copilot. The hosts look at global trends, with a report showing accelerated AI adoption in Latin America despite challenges. They also explore new products, as Perplexity makes its AI-powered Comet browser free for everyone. They conclude with geopolitical and social issues, as a US report flags Chinese AI models as a security risk, and the use of AI in religion causes division. Hosted on Acast. See acast.com/privacy for more information.