Latest / Neural Newscast / AI Rivals Sued for Coordinating a Slowdown Agreement [Model Behavior]
Transcript
- [00:00] Thatcher Collins: From Neural Newscast, this is Model Behavior, [00:03] Thatcher Collins: AI-focused news and analysis on the models shaping our world. [00:12] Nina Park: I'm Nina Park. Welcome to Model Behavior, a program from Neural Newscast that examines [00:17] Nina Park: how artificial intelligence systems are built, deployed, and operated in professional environments. [00:24] Nina Park: Our top story this Monday morning involves a significant legal challenge [00:27] Nina Park: to the coordinated safety efforts of the world's leading AI laboratories. This past Friday, [00:33] Nina Park: a class action lawsuit was filed in California, targeting [00:36] Nina Park: the specific pace of model development and the structure of industry-wide safety agreements. [00:42] Thatcher Collins: I'm Thatcher Collins. Nina, this lawsuit filed in the United States District Court for [00:47] Thatcher Collins: the Northern District of California targets the industry's four largest players: Anthropic, OpenAI, [00:53] Thatcher Collins: SpaceX AI, and Google. The plaintiffs represent paid subscribers to services like ChatGPT, Claude, [01:00] Thatcher Collins: Grok, and Gemini. They claim that an agreement between these chief rivals to intentionally slow [01:06] Thatcher Collins: their development pace constitutes a violation of antitrust laws, [01:10] Thatcher Collins: effectively reducing the technological value that consumers receive for their subscriptions. [01:15] Nina Park: The timeline here is quite specific. According to the filing, the coordination became public on [01:20] Nina Park: September 12th. That was when Anthropic CEO Dario Amodei published an essay urging for [01:26] Nina Park: unified industry-wide cooperation on decelerating advancements to favor long-term safety measures. [01:33] Nina Park: Almost immediately, Sam Altman, Elon Musk, and Demis Hassabis all publicly signaled their agreement [01:39] Nina Park: with the proposal. Thatcher, how are the plaintiffs [01:42] Nina Park: characterizing this public alignment between rivals who are normally in such fierce competition? [01:47] Thatcher Collins: They see this coordination as a legal shortcut. Lead attorney Nick Rowley argues [01:52] Thatcher Collins: that while individual companies are free to slow their own progress, they cannot legally substitute [01:57] Thatcher Collins: collective restraint for individual accountability. The defendants, however, [02:01] Thatcher Collins: point to existential risks. Amadee warned that without a coordinated slowdown, [02:06] Thatcher Collins: AI systems could be capable within six to twelve months of directing a swarm of agents [02:11] Thatcher Collins: that could take over critical internet infrastructure. [02:14] Thatcher Collins: They believe the competitive pressure is exactly why they need a shared agreement. [02:18] Nina Park: It is interesting [02:19] Nina Park: because Amade actually acknowledged these potential antitrust hurdles in his essay. [02:25] Thatcher Collins: He suggested that the United States [02:26] Thatcher Collins: government should mediate these discussions or provide a narrow legal waiver [02:31] Nina Park: for these safety conversations. [02:33] Thatcher Collins: But Thatcher, the plaintiffs argue that allowing private for-profit companies to control [02:38] Thatcher Collins: these safety protocols through self-serving agreements is exactly what puts the public [02:43] Nina Park: at risk. They want to see these technical standards [02:46] Nina Park: established through transparent regulation, rather than opaque backroom handshakes. [02:51] Thatcher Collins: And that is where we run into the current political reality. President Trump has vocalized [02:56] Thatcher Collins: his rejection of these calls for regulation, claiming they are part of a conspiracy to slow [03:01] Thatcher Collins: American progress. He specifically questioned why industry leaders would support rules [03:06] Thatcher Collins: that he claims could drive their own firms into bankruptcy. The administration's priority [03:11] Thatcher Collins: remains ensuring American labs outpace Chinese competition, viewing any force slowdown [03:16] Thatcher Collins: as a strategic disadvantage. Over the weekend, [03:19] Thatcher Collins: Trump even announced plans for a dedicated AI task force and czar. [03:23] Nina Park: The lawsuit also points to a statement from July 2026 as evidence that this coordination [03:29] Nina Park: has been brewing for months. At that time, high-ranking employees across several labs signed [03:35] Nina Park: a document acknowledging the intense pressure to keep moving at breakneck speed. Thatcher, [03:40] Nina Park: this highlights a significant paradox: if these companies genuinely believe the technology [03:45] Nina Park: is becoming dangerous, but federal antitrust laws prevent them from agreeing to stop, [03:50] Nina Park: we are left in a very precarious regulatory gray area where safety and competition land conflict. [03:57] Thatcher Collins: Exactly, Nina. Senator Josh Hawley echoed this concern in a recent hearing, [04:01] Thatcher Collins: stating he would never support an antitrust exemption for the world's most powerful companies. [04:06] Thatcher Collins: He argues it would just become a tool to stifle smaller competitors. So the labs are stuck between [04:11] Thatcher Collins: a lawsuit claiming they are colluding and a government that insists they keep accelerating. [04:16] Thatcher Collins: For the paid subscribers bringing the suit, the issue is simpler. They are paying for [04:20] Thatcher Collins: the most advanced tech possible and feel they are being denied it by a cartel-like agreement. [04:26] Nina Park: This case will likely set the precedent for whether safety is a valid defense [04:30] Nina Park: for market coordination in the AI era. We will continue to track the responses from Anthropic, [04:37] Nina Park: OpenAI, and the other defendants as they become available. Thank you for listening to [04:42] Nina Park: Model Behavior, a Neural Newscast editorial segment. For further details on this litigation [04:49] Nina Park: and our full archive, visit mb.neuralnewscast.com. Neural Newscast is AI-assisted, [04:56] Nina Park: human-reviewed. View our AI transparency policy at neuralnewscast.com. [05:02] Announcer: This has been Model Behavior on Neural Newscast, examining the systems behind the story.