{
  "id": 3675402,
  "title": "The least bad way to regulate AI?",
  "url": "https://urgent.news/2026/08/27/the-least-bad-way-to-regulate-ai",
  "topic": "ai",
  "section": "AI",
  "published": "2026-08-27T05:03:47.000Z",
  "source": {
    "name": "Marginal Revolution",
    "slug": "marginal-revolution",
    "url": "https://marginalrevolution.com/marginalrevolution/2026/08/the-least-bad-way-to-regulate-ai.html"
  },
  "original_language": "en",
  "account": "The column explores the question of how best to regulate AI technologies. The author argues that some basic safeguards should be put in place, but without hampering overall AI progress. It suggests trusting the AI labs themselves to act as primary regulators, rather than following conventional wisdom about public oversight.\n\nA key part of the proposal is to establish a private not-for-profit body dedicated to AI regulation. Inspired by FINRA's model, this consortium of AI firms would examine trade practices, make recommendations, and help the federal government with oversight and regulation. Major AI labs would be part of this body and would be overseen by Washington, potentially through the Center for AI Standards and Innovation.\n\nThe body would periodically audit major AI companies and their models, assessing their conduct and safety. At first, much of this would focus on cybersecurity, and whether new models create more cyber risk than they help to solve. If a company passed the audit, it would be exempt from standard liability law, as long as it demonstrated reasonable care. This would remove fear of large liability awards and give AI labs strong incentives to meet safety standards.\n\nThe author acknowledges that such a body composed of industry players might not always issue fair judgments. However, there are many benefits and no better alternative. Companies know that a dangerous model from another firm could lead to a harmful incident and damage the entire industry's prospects, similar to the Three Mile Island accident of 1979. Competitors also want to avoid too much competition from lower-price, lower-quality startups, which may lead them to support tough safety standards. Ultimately, we are forced to choose between imperfect options, with the reality being that there is far more expertise within the companies to judge AI safety than the federal government can realistically provide anytime soon.",
  "summary": "That is the topic of my latest Free Press column. Excerpt: The key is to create some basic safeguards, but without stifling broader AI progress. To do so, we must defy the conventional wisdom about public oversight and instead trust the AI labs to be their own primary regulators. My version of the proposal starts […] The post The least bad way to regulate AI? appeared first on Marginal REVOLUTION…",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}