AI Summary
5 min readChinese AI models have recently rattled Washington and Silicon Valley by matching the performance of top U.S. models like those from OpenAI and Anthropic—at a fraction of the cost. The Trump administration is now divided over whether to restrict American companies from using them, with national security officials warning they could be weaponized for cyberattacks or bioweapons, while AI advisors argue that open access is key to defense.
The New Batch of Chinese AI Models
Chinese companies, including Kimi (from Moonshot AI) and Alibaba’s Qwen, have released open-source AI models that industry benchmarks suggest are competitive with the most advanced U.S. models. These models are significantly cheaper to use and lack many of the security guardrails that top U.S. models have. This combination has excited investors and analysts, but it has also alarmed policymakers. As WSJ reporter Amrith Ramkumar notes, the U.S. companies are “rattled” because they are essentially asking, “Why are you letting these Chinese companies do things that we can’t?” The dynamic represents “another reset in the AI race.”
Security Concerns: Two Layers
Continue reading the full summary in the app — free to try.
Read Full Summary →Free • No credit card required
Never miss an episode of WSJ What’s News
Get every new episode summarized in your inbox — free, ~5 minutes to read.
No spam. Unsubscribe anytime.
What you'll learn
- 1 (07:42) **Chinese AI Models Rattle Silicon Valley** - Chinese companies release cheaper, open-source AI models that rival top US models, sparking alarm in Washington and Silicon Valley.
- 2 (08:18) **How Chinese Models Stack Up** - Early excitement over the models' performance and cost, but definitive benchmarks are not yet available.
- 3 (09:12) **Administration Divide on Restricting Chinese AI** - Officials focused on national security want to restrict access, while AI advisors like David Sachs advocate for keeping tools accessible for defense.
- 4 (09:45) **Two Layers of Security Concern** - The classic data privacy issue and the bigger worry that open models could be used to launch cyber attacks or develop new pathogens.
- 5 Standout Quotes
- 6 (10:04) “The concern is that you have this open model that is capable... Almost anyone can use it. So that is really setting off alarm bells for a lot of policy makers. If you end up in that scenario, and it seems like we could potentially be there very soon, if not all...” — WSJ Reporter Amrith Ramkumar
- 7 Mentioned Resources
+ Full timestamped outline available in the app
Guests on this episode
Show Notes
P.M. Edition for July 21. Recent AI models from China claim they’re just as powerful as some of the most cutting-edge models from OpenAI and Anthropic. Journal reporter Amrith Ramkumar joins to discuss the latest reactions from Silicon Valley and the White House. Plus, General Motors had a strong second quarter as consumers kept buying pickup trucks and SUVs. And New Jersey says a software error led to almost 400 non-citizens voting in elections in the state since 2023. Alex Ossola hosts.
Sign up for the WSJ's free What's News newsletter.
Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.
More from this podcast
WSJ What’s News →