When does SEO optimization for LLM model outputs begin? How much value is there in paying off RLHF workers to promote your brand? Robert Mahari

Benchmarks and risk assessments don't mean much when you can't be sure what model you're really using. Here we demonstrate a cryptographic approach to verifiable model attestations -- an exciting example of Regulation by Design for AI! Paper: arxiv.org/pdf/2402.02675…

What happens when we fall in love with our robots? In the MIT Technology Review, Pat Pataranutaporn and I discuss the challenges of regulating AI companions and explore new regulatory approaches that address the unique harms of addictive intelligence. technologyreview.com/2024/08/05/109…