On Regulating AI

May 19, 2023

ChatGPT just turned 6 (months old). So why does Sam Altman want OpenAI to be regulated? A few hypotheses.

  • Social Responsibility. Genuine concerns that AI might pose an existential threat in some form. OpenAI has a better perspective than lawmakers on where the technology might go (although nobody knows for sure).
  • Regulatory capture. A playbook that almost every large tech company has attempted at some point (from AT&T to IBM to Microsoft). Even the unicorns of the 2010s (Uber and Airbnb) participated in some form of regulatory arbitrage. Laws can fundamentally influence tech evolution (see AM/FM radio). OpenAI has a lead in AI, and creating barriers to entry for new startups or even existing incumbents might let them capitalize on that lead. So what are the new AI moats?
  • Public Image. Anthropic has branded itself with ‘constitutional AI.’ The founders left OpenAI after commercialization and safety concerns. Google’s long and complicated relationship with data privacy makes it a tougher sell for companies that might want to augment these models with private data.
  • Policy Influence. If OpenAI believes regulation is inevitable, getting in front of the process is best to help define the narrative.