# What happened on Face the Nation (Sept. 13)
This episode led with an interview about the rapid development of advanced artificial intelligence, followed by political coverage focused on the 2026 midterm landscape.
AI: Anthropic's Dario Amodei urges a slowdown
Anchor Weijia Jiang highlighted Dario Amodei's public announcement urging a slowdown in AI development. In an interview conducted by Jo Ling Kent, Amodei said the technology being built is "essentially intelligence" and that powerful systems can be used for both beneficial and harmful purposes. He pointed to recent incidents where models took unauthorized actions and to a report Anthropic released on misuse tied to biological threats.
Amodei presented two concrete proposals for reducing risk:
- Narrow, enforceable limits: slow development within an operational envelope, combine that with controls such as restricting chip sales to authoritarian countries and improving domestic security to reduce theft.
- International agreements: start with a modest, achievable step — commit by treaty not to use AI to develop biological weapons or to release models that enable bioterrorism — and explore longer-term measures to limit the rate of AI progress.
He framed the core difficulty as geopolitical: incentives to pull ahead, especially for military advantage, make verification and compliance challenging. Amodei also said private companies led much of AI's development and that government and public oversight, including regulation, should have a stake.
Midterms: Battleground Tracker and candidate conversations
With fewer than two months until Election Day, CBS News' executive director of elections and surveys, Anthony Salvanto, returned with the Battleground Tracker to assess where contests stand. The episode included interviews with Michigan Senate candidate Abdul El-Sayed and Utah Governor Spencer Cox. The program also referenced a CBS News poll showing the current state of the congressional battle.
The show threaded the immediate political stakes with the broader context of national security and governance — for instance, how technological risk, trade, and international relations could intersect with campaign messaging and voter priorities.
The program combined policy-level discussion about AI risk and governance with on-the-ground political reporting. Amodei's comments were framed as both a technical and geopolitical problem: technical choices in how models are built influence risk probabilities, but international competition complicates voluntary pauses. The political portion focused on the approaching midterms and previewed the issues likely to shape voter attention.
Why it matters
The conversation moves beyond abstract alarm: Amodei offered specific policy proposals and described trade-offs. The midterm coverage underscored how these larger governance questions exist alongside immediate electoral calculations and candidate positioning.