The United States is considering sanctions against Chinese open-source AI models. Treasury Secretary Scott Bessent has indicated that the Trump administration might act against Chinese AI companies accused of stealing American intellectual property to train their systems.
What’s Actually Being Threatened
Bessent’s warning marks a clear escalation in the ongoing tech rivalry between the US and China. The focus is on Chinese open-source AI models — systems with publicly available code that are trickier to ban than closed commercial products that can simply be removed from app stores.
The core accusation is that Chinese AI developers have used copyrighted American data, code, or model outputs to train their AI without permission. Imagine photocopying an entire textbook and claiming it as your own — except here, the “textbook” is worth billions in AI research.
If sanctions happen, American companies and individuals might be legally barred from using, downloading, or engaging with sanctioned Chinese AI models. This could be a major issue for developers relying on tools like DeepSeek, a Chinese open-source AI model that gained attention earlier in 2026.
Why Open-Source AI Is a Tricky Target
Sanctioning a physical product like a chip or a company’s bank accounts is pretty straightforward. But sanctioning software that’s already been downloaded millions of times is much more complicated.
Once an open-source model hits the internet, it spreads quickly. Copies exist on personal computers, private servers, and cloud platforms around the world. While a US sanction could stop American entities from accessing new versions or working with developers, it can’t erase the copies already in use.
This is why the move is being labeled as a threat for now — the legal and logistical details of enforcing such a sanction are still being figured out.
A Second Front: Apps on US Military Devices
The sanction threat isn’t the only AI-related security worry surfacing this week. A report from Ars Technica found that over one in eight apps studied contained code from Chinese or Russian sources — and some of these apps are used by US military personnel.
This embedded foreign code can collect data, track locations, or create backdoors that give outsiders access to systems. The implications for national security are clear, especially when those users are American soldiers.
These two stories highlight the US government’s growing concern about the origins of AI and software, not just their functionalities.
The Bigger Picture: Slowing China’s AI Race
This latest action fits into a broader strategy. The Trump administration has already limited sales of advanced AI chips to China, added Chinese tech firms to trade blacklists, and urged allies to restrict their technology transfers. Targeting AI models is the next logical step.
Washington is worried that Chinese AI developers are catching up to American companies faster than anticipated, partly due to shortcuts involving American intellectual property.
| By The Numbers | |
|---|---|
| Share of analyzed apps containing foreign code | More than 1 in 8 (12.5%+) |
| Apps flagged as targeted at US military personnel | Multiple, per Ars Technica report |
| Primary enforcement tool being considered | Treasury Department sanctions |
| US official making the threat | Treasury Secretary Scott Bessent |
What This Means for Everyday Users
If you’re a regular consumer, the immediate impact of a sanction threat is minor — you probably aren’t downloading raw AI models from Chinese research labs. But the ripple effects could be significant.
Many apps and services — like chatbots, translation tools, and image generators — rely on open-source AI models. If developers have to remove that underlying technology, you might see apps malfunction, slow down, or vanish from stores without explanation.
For businesses using Chinese AI tools to save costs, a formal sanction would force them to switch to alternatives. This transition would likely mean turning to American models from OpenAI, Google, or Meta.
The story about apps with foreign code is more personal. If you use apps targeted at military communities, fitness tracking, or communication — common among service members — it’s worth checking if those apps have faced security scrutiny.
Community Reaction
“The open source angle is what makes this so complicated. You can’t un-publish a model. Sanctions would just push usage underground or overseas.”
“Military apps with Chinese code is the one that should be getting more attention. That’s not a trade dispute, that’s a security failure.”
What To Watch
- Treasury Department guidance: Keep an eye out for any formal rules or executive orders from Treasury that might turn Bessent’s threat into enforceable policy. This process usually takes weeks to months.
- DeepSeek and similar models: Chinese open-source AI models currently in use among American developers will be a key test — watch for statements from major cloud providers like AWS and Google Cloud about their hosting plans for these models.
- Congressional response: Several bills aimed at Chinese AI and software are already moving through Congress. The timing of Bessent’s comments suggests the executive branch may be aligning with legislative efforts.
- App store policy changes: If the military app findings gain traction, Apple and Google may come under pressure to audit foreign code in apps available on their platforms — especially those used in sensitive government or defense contexts.
Sources
Maya Torres
Maya Torres is the Consumer Tech Editor at Explosion.com with 7 years covering product launches for major technology publications. She has reviewed over 300 devices across smartphones, laptops, wearables, and smart home products. Maya specializes in translating spec sheets into real-world buying advice and attends CES, MWC, and Apple keynotes as press. Her reviews focus on helping readers decide what to buy, not just what specs look good on paper.
