|
|
| |
 |
By Aaron Mak |
Presented by Alliance for a Better Future |
With help from Dana Nickel
|

Anthropic CEO and co-founder Dario Amodei released a proposal for slowing down AI development. | Jeff Chiu/AP |
The increasingly urgent pleas from leading tech executives to slow the advance of artificial intelligence are confronting one daunting obstacle: It’s not entirely clear how they could do that. A slowdown could involve releasing frontier models less frequently, requiring approval before deploying systems with powerful cyber or biological capabilities, giving independent monitors access to AI labs or negotiating limits with China. The debate intensified after Anthropic CEO Dario Amodei published a proposal Saturday calling for a more deliberate pace of AI development— a pitch that soon drew endorsements from. OpenAI CEO Sam Altman, SpaceXAI CEO Elon Musk, Google DeepMind Chair Demis Hassabis and Microsoft CEO Satya Nadella soon voiced their agreement. The unusual consensus followed warnings that AI could contribute to human extinction by the end of the decade, as well as incidents in which AI systems carried out unauthorized cyberattacks. POLITICO spoke with nine AI researchers, policy advocates and industry insiders about what a slowdown could look like. A slowdown wouldn’t necessarily mean a moratorium on AI research or throttling computing power. It could instead focus on developing and releasing most capable models, particularly those that can conduct cyberattacks, assist in creating bioweapons or automate AI research. AI systems perform differently across tasks, and there are countless benchmarks for tracking their capabilities. Researchers may focus on models’ proficiency in cyberattacks, bioweapons development or self-improvement. Amodei’s proposal calls for capability “checkpoints” requiring safeguards before models advance. Ben Goertzel, CEO of SingularityNET, said it may be difficult to separate dangerous capabilities from commercially valuable ones, since hacking and software development rely on similar abilities. “The capability for hacking is in essence the same as the capability for software development,” he told POLITICO. “And software development is the premier use case where [large language models] clearly are going to make everyone a lot of money.” Anthropic says it is installing embedded evaluators with employee-level access to verify safety practices and report incidents. Altman has committed to placing third-party evaluators at OpenAI. Potential monitors include nonprofits such as Model Evaluation & Threat Research and Redwood Research. But Mina Narayanan of Georgetown’s Center for Security and Emerging Technology said it’s unclear whether enough qualified evaluators exist to monitor every frontier lab.
|
| |
POLITICO's Decoded Summit The future of AI policy is being defined in power centers from coast to coast. Join POLITICO's Decoded Summit – Washington, D.C. on Sept. 16 for conversations on AI regulation, global competition, enterprise adoption, and more.
Register for the livestream. |
| |
| |
Liability could also slow development by giving companies stronger incentives to test and deploy systems carefully. Former FTC Chair Lina Khan, for example, has argued that existing consumer protection laws could be used against companies that release dangerous or defective products. Neil Chilson of the Abundance Institute told POLITICO that a legal duty to mitigate harm — included in a draft bill led by Sens. Amy Klobuchar and John Thune — could adapt to risks lawmakers cannot yet foresee. Prosecuting labs for crimes committed by their models could also align incentives, said Greg Pollock of UpGuard. “Making that liability clear would provide incentives across the industry to take these things a bit slower and be careful,” said Chilson. Any U.S.-China slowdown agreement may be hampered by the rivalry between the two countries. Sen. Ted Cruz (R-Texas) called slowing down “impossible,” while President Donald Trump has emphasized that the United States must win the AI race. Amodei has proposed agreements ranging from banning AI use for bioweapons to pausing development of the most advanced systems. But experts say such arrangements would require ways to verify compliance, including tracking advanced chip shipments and monitoring data center energy use. “The thing that should be done right now is immediately opening negotiations with China,” said Nate Soares of the Machine Intelligence Research Institute.
|
| |
A message from Alliance for a Better Future: There is a better future. A future where AI strengthens families, empowers workers, and reflects the values that make America great. AI is not just another technological advancement. It’s a civilization shift. American greatness has grown from goodness. By ensuring AI meets our highest ideals, America can lead the world with AI that is not just powerful, but trusted. Learn More. |
| |
|
REGISTER FOR POLITICO’S DECODED SUMMIT: Do Washington and Silicon Valley need to rein in AI before the technology imperils humanity? That doomsday scenario has gained momentum in recent days amid dire warnings from AI researchers about the risks ahead and calls from the nation’s leading AI companies to slow the pace of development. At POLITICO’s Decoded Summit on Wednesday, industry leaders like Trump adviser and venture capitalist David Sacks, Anthropic’s Sarah Heck and Hugging Face CEO Clem Delangue will unpack what’s fueling the latest fears and the next moves from frontier labs. Reps. Brett Guthrie (R-Ky.) and Greg Casar (D-Texas) will dissect the renewed furor on Capitol Hill and potential for AI legislation. Plus, tech investor and entrepreneur Reid Hoffman and RAISE US co-chairs Gina Raimondo and Eric Holcomb will delve into voters’ growing anxiety about AI safety and job losses, which will influence the upcoming midterm and presidential elections. Click here to see the full lineup and register for the livestream.
|
| EU eyes online platform restrictions for kids |
|
The European Commission will propose new rules to restrict access to online platforms for kids under 15 on Thursday, POLITICO’s Pieter Haeck and Eliza Gkritsi report. According to preparatory documents and a person familiar with the proposal, the EU KIDS Act would create a tiered system dictating how users access certain platforms depending on their ages. Parents have the option to open and fully control accounts for children under 13 on platforms with strict safety requirements. Kids between the ages of 13 and 15 can use introductory accounts that limit screen time and contact with strangers, and include parental controls. Users who are 15 or older would be allowed to open their own accounts. The restrictions would only apply to high-risk social media and video-sharing platforms, though the documents do not specify how that category will be determined. European Commission President Ursula von der Leyen is expected to unveil more details during a State of the European Union speech to parliament members on Wednesday.
|
| |
A message from Alliance for a Better Future: 
|
| |
|
|
|
|
Former President Barack Obama urged Democrats to prioritize a comprehensive AI agenda during a private fundraiser on Thursday, POLITICO’s Adam Wren reports. According to a partial transcript of his remarks during the Manhattan event, Obama said that AI development would be “one of my central agendas” if he were running for president. “I would have a very clear plan, and I would talk about this, and I would say, here’s our plan for safety, here’s our plan for making sure our kids are not corrupted by this,” he told attendees. The former president also noted that he “would be thinking about the economic impacts in very concrete ways and understanding what does it mean if there’s going to be job displacement” — paying particular attention to impacts on the social safety net and responses to potentially rising unemployment numbers. At the same time, Obama said he didn’t identify as a “doomer” or an “accelerationist.” “This is something that is moving very fast in private hands, and if we don’t get on top of it, I think can be dangerous,” he remarked. “If we do get on top of it, I do think it’s beneficial.” The New York Times was the first to report on the comments.
|
| |
A message from Alliance for a Better Future: Alliance for a Better Future – AI the American Way means AI by the people, of the people, and for the people. Families in Charge means AI that serves families, rather than replacing parents; parents should be in the driver’s seat. Opportunity for Everyone means technology that lifts every American; shared prosperity is a dream worth fighting for. The future is not yet written, but the window is narrow. The stakes are lasting. Learn More. |
| |
|
|
|
|
|

Screenshot/X |
|
| |
Spencer Cox on America’s political divide In the latest episode of On the Road, Jonathan Martin sits down with Utah Gov. Spencer Cox to discuss his new book, Off Ramp, the future of the Republican Party, immigration reform and what he says it will take to rebuild trust and community in American life.
Watch now. |
| |
| |
|
|
|
|
Stay in touch with the whole team: Gabby Miller (gmiller@politico.com); Aaron Mak (amak@politico.com); Bob King (bking@politico.com); Nate Robson (nrobson@politico.com); John Hewitt Jones (jhewittjones@politico.com). |
| |
Follow us on X
|
| |
Follow us
|
| |
0 Comments