The Lawfare Podcast Scaling Laws: AI Safety Meet Trust & Safety with Ravi Iyer and David Sullivan
10 snips
Oct 10, 2025 
David Sullivan, Founding Executive Director of the Digital Trust & Safety Partnership, and Ravi Iyer, Managing Director at USC’s Psychology of Technology Institute, delve into the evolving landscape of Trust and Safety in the digital realm. They discuss the need for user-centered regulations, the role of AI in moderation, and the importance of design over top-down policies. The duo explores how AI companions differ from social media in purpose and agency, and what best practices can be applied to ensure safer digital experiences, particularly for vulnerable users like teens.
AI Snips
Chapters
Transcript
Episode notes
Design Upstream To Prevent Harm
- Use design interventions like privacy-by-default, rate limits, and youth protections to prevent harms upstream.
- Change recommendation and visibility systems to stop risky interactions before they start.
Layoffs Didn't Erase Expertise
- Layoffs in trust and safety teams reflect both political pushback and recognition that hiring alone won't fix systemic problems.
- Many laid-off T&S staff dispersed to smaller platforms, vendors, academia, and AI solutions.
Measure Harm By Asking Users
- Platform policy metrics (e.g., percent content violating policy) poorly map to real user harm.
- Measuring harm requires asking users directly via surveys and tracking experience indicators over time.