Skip to content
Source · Daily Brief

AI Daily Brief — 16 January 2026

Friday’s two threads. Stanford publishes a study showing production LLMs reproduce large portions of copyrighted books — Claude 3.7 Sonnet recalls 95.8% of Harry Potter and the Sorcerer’s Stone nearly verbatim; GPT-4.1, Gemini 2.5 Pro and Grok 3 show similar patterns. Microsoft reframes its search guidance around Answer Engine Optimization (AEO) and Generative Engine Optimization (GEO): AI systems now judge products by data completeness, structured markup and currency rather than legacy keyword tactics.

Top stories

  • Stanford — production LLMs leak copyrighted books. Claude 3.7 Sonnet at 95.8% verbatim on Harry Potter and the Sorcerer’s Stone; GPT-4.1, Gemini 2.5 Pro, Grok 3 all show similar reproduction patterns. via MarketingProfs
  • Microsoft — AEO + GEO replace SEO playbook. AI systems prioritise data completeness, structured markup and currency. via MarketingProfs

Backdrop

Google Veo 3.1 (released 13 January) continues to land in production-quality reviews. Anthropic‘s mid-week leadership reorg — Mike Krieger moving from CPO to co-lead the Labs incubator with Ben Mann, Ami Vora taking Head of Product — frames the rest of January.

Compiled by AI Feed’s editor from publicly reported announcements on 16 January 2026.