Can political institutions make competent AI regulatory decisions today?

Vince Hankes · May 3, 2023

Regulation of AI is coming and is necessary, but it must be designed in partnership with the companies building the technology; poorly designed regulation this early would be ineffective for everyone

Regulation for the sake of regulation won't solve the problems people are concerned about — you have to understand the technology's nuances, so builders must help educate regulators

Scope: speaking personally, not for OpenAI

30:33 20VC: The OpenAI Memo: Why Invest? Is it too Late to Catch OpenAI? Are OpenAI's Models Truly Defensible? Does the Value in AI Accrue to Incumbemts or Startups - Application Layer/Infrastructure? What Happens with Regulation? with Vince Hankes @ Thrive

Vince Hankes · May 3, 2023

The knowledge gap between AI builders and regulators is a solvable problem

The information and experts exist, and regulators got impressively up to speed on crypto by taking time to learn; the dialogue between AI practitioners and regulators just has to start

Scope: may take years; regulators will still make mistakes along the way

31:55 20VC: The OpenAI Memo: Why Invest? Is it too Late to Catch OpenAI? Are OpenAI's Models Truly Defensible? Does the Value in AI Accrue to Incumbemts or Startups - Application Layer/Infrastructure? What Happens with Regulation? with Vince Hankes @ Thrive

Alex Lebrun · Jun 19, 2023

The new EU AI regulation is a disaster because it was written too fast and without involving the people actually building in the domain

Good regulation depends on timing and on involving practitioners; Europe, being late technologically, wanted to at least be first at regulating

43:28 20VC: Why No Models Today Will Be Used in a Year, Why Open Will Always Beat Closed in AI, Why Proprietary Data is Less Important Than Ever And Why EU AI Regulation is a Disaster with Alex Lebrun, Founder & CEO @ Nabla

Reid Hoffman · Jun 10, 2024

The political system can make good AI regulatory decisions when the right smart people are involved — Macron, the White House executive order, and the UK AI Safety Institute are all doing well

Existing examples of competent policy show it can be done

Scope: conditional on talking to the right smart people

47:24 20VC: Reid Hoffman on Foundation Models: Who Wins & How Do Incumbents Respond | The Inflection AI Deal: How it Went Down | Why Trump is a Threat to Democracy | The Future of TikTok | Lessons from Sam Altman, Brian Chesky and the OpenAI Board

Harry Stebbings · Apr 28, 2023

The gap between regulators' knowledge and the actual technology has never been bigger, leaving regulatory bodies unable to regulate AI with domain knowledge

7:25 20VC: In AI Who Wins? Startups or Incumbents? What Happens to Wealth Inequality? Why Will $10BN+ Companies Only Have 10 People | Why Defensibility in Startups is BS & Speed is Everything? Why Large Groups Worsen Decision-Making with Sarah Guo

Harry Stebbings · May 3, 2023

The knowledge gap between the private sector and regulators has never been greater, which makes effective AI regulation hard to set

Regulators do not understand much of the infrastructure or the opportunity

31:39 20VC: The OpenAI Memo: Why Invest? Is it too Late to Catch OpenAI? Are OpenAI's Models Truly Defensible? Does the Value in AI Accrue to Incumbemts or Startups - Application Layer/Infrastructure? What Happens with Regulation? with Vince Hankes @ Thrive

Sarah Guo · Apr 28, 2023

The regulator knowledge gap in AI is a real concern but is not structurally different from other technology areas like internet cybersecurity; what is genuinely different is the speed of change

Every technology wave has produced similar policy gaps, but here society does not have decades to adjust to new capabilities

7:40 Knowledge gap is not new but speed of ai change is unprecedented

Alex Lebrun · Jun 19, 2023

AI regulation should be delayed by a year or two and then written with input from model builders, users and the general public rather than only experts

We are still in the demo phase of AI — LLMs are barely used — so nobody yet knows what the real risks are as opposed to science-fiction risks; you need field experience to know what to regulate and why

44:54 Regulation should wait for field experience and broad stakeholder input not just experts