The TED AI Show What really went down at OpenAI and the future of regulation w/ Helen Toner
100 snips
May 28, 2024 
Former OpenAI board member and AI policy expert, Helen Toner, discusses internal conflicts at OpenAI, deceptive practices, regulating AI in surveillance technologies, combating AI-driven voice and video scams, and the complexities of regulating AI technology and policymaking.
AI Snips
Chapters
Transcript
Episode notes
Surveillance Is A Quantitative Shift, Not Just Qualitative
- AI surveillance scales existing visibility into ubiquitous, automated monitoring with new biases.
- Widespread deployment changes social dynamics and raises serious privacy and due-process risks.
Ditch Voice Authentication For Security
- Be cautious about voice-based authentication and warn relatives about voice-deepfake scams.
- Prefer stronger, non-voice security methods for banks and sensitive services.
Transparency Beats Blanket Bans
- Transparency requirements let government track advanced model development without banning innovation.
- Reporting and measurement help policymakers respond to rapidly changing AI capabilities.