US Seeks To Ban 'Adversarial AI'

A cross-party group of U.S. lawmakers have introduced the “No Adversarial AI Act,” a bill designed to prohibit federal agencies from using artificial intelligence models developed by companies based in China, Russia, Iran, or North Korea. 

The bill mandates that the Federal Acquisition Security Council maintain a regularly updated list of banned AI technologies, with China’s DeepSeek cited as a prime example of the threat.

The legislation, sponsored by both Democrat and Republicans,  responds to growing concerns about the national security risks posed by foreign AI systems, particularly those with ties to those considered to be adversarial governments.

At the heart of the legislation is DeepSeek, a Chinese AI startup that has drawn scrutiny for its alleged connections to the Chinese Communist Party (CCP) and its intelligence apparatus. A prior investigation by members of the House of Repersenetaives revealed that DeepSeek’s privacy policy explicitly states that U.S. user data is stored in China, raising fears that sensitive information could be accessed by the CCP.

Reports also suggest that DeepSeek has been aiding China’s military and intelligence operations, including through the use of Southeast Asian shell companies to procure high-end semiconductors restricted by U.S. export controls.

These findings have intensified calls for a federal procurement ban on AI models originating from China, a recommendation now embodied in the proposed legislation.

Privacy & Security Concerns

The “No Adversarial AI Act” reflects deep concerns about data privacy and the potential for foreign AI systems to undermine U.S. security. Lawmakers argue that AI models like DeepSeek, which reportedly incorporate code capable of transmitting user data to a sanctioned Chinese state-owned telecom, pose a direct threat to federal systems and sensitive government data. “The U.S. government should not be sending our data to China,” Democrat Representative Raja Krishnamoorthi said during a hearing of the Select Committee on the CCP, emphasising the need for a “clear firewall” between U.S. agencies and adversarial AI technologies. 

A Broader Geopolitical Context

The bill arrives amid escalating tensions between the U.S. and China over technological dominance. Critics of DeepSeek argue that its open-source AI models, which can be installed locally, are being unfairly targeted to protect U.S. Big Tech monopolies. However, proponents of the legislation counter that the risks outweigh any competitive concerns, particularly given DeepSeek’s alleged role in supporting Beijing’s military ambitions. The act also includes provisions for removing AI models from the banned list if they are proven free of adversarial influence, offering a potential path for compliance but placing the burden of proof on the companies.

The “No Adversarial AI Act” represents a significant step in Washington’s efforts to safeguard national security in the AI era.

While the bill enjoys bipartisan support, its passage is not guaranteed, as debates over technological competition and the feasibility of enforcement continue. For now, the legislation sends a clear message: the U.S. is determined to limit the influence of adversarial AI in its government systems, with DeepSeek squarely in the crosshairs.

As Krishnamoorthi stated, “Artificial intelligence controlled by foreign adversaries poses a direct threat to our national security, our data, and our government operations.”

Stocktwits  |   Economic Times  |  Cyber Express  |   Reuters  |  APNews  |   Benjamin Norton 

Image: Ideogram

You Might Also Read: 

DeepSeek - A Deep Dive Reveals More Than One Red Flag:


If you like this website and use the comprehensive 8,000-plus service supplier Directory, you can get unrestricted access, including the exclusive in-depth Directors Report series, by signing up for a Premium Subscription.

  • Individual £5 per month or £50 per year. Sign Up
  • Multi-User, Corporate & Library Accounts Available on Request

Cyber Security Intelligence: Captured Organised & Accessible


 

« Notorious Hacker ‘IntelBroker’ Arrested
The UK’s Software Security Code Of Practice Is More Than Just A Guidance »

ManageEngine
CyberSecurity Jobsite
Check Point

Directory of Suppliers

Clayden Law

Clayden Law

Clayden Law advise global businesses that buy and sell technology products and services. We are experts in information technology, data privacy and cybersecurity law.

BackupVault

BackupVault

BackupVault is a leading provider of automatic cloud backup and critical data protection against ransomware, insider attacks and hackers for businesses and organisations worldwide.

IT Governance

IT Governance

IT Governance is a leading global provider of information security solutions. Download our free guide and find out how ISO 27001 can help protect your organisation's information.

Authentic8

Authentic8

Authentic8 transforms how organizations secure and control the use of the web with Silo, its patented cloud browser.

NordLayer

NordLayer

NordLayer is an adaptive network access security solution for modern businesses — from the world’s most trusted cybersecurity brand, Nord Security. 

BSI Group

BSI Group

BSI is the business standards company that equips businesses with the necessary solutions to turn standards of best practice into habits of excellence

MaxMind

MaxMind

MaxMind is an industry-leading provider of IP intelligence and online fraud detection tools.

4N6

4N6

4N6 is a privately-owned firm founded with the goal of providing expert knowledge of computer forensics.

Trulioo

Trulioo

Trulioo is a leading global identity and business verification company providing secure access to data sources worldwide to instantly verify consumers and businesses online.

CyberGuru

CyberGuru

CyberGuru is a service provided by CyberSecurity Malaysia specializing in cyber security professional training and development.

MISP Project

MISP Project

The MISP threat sharing platform is a free and open source software helping information sharing of threat intelligence including cyber security indicators.

Splone

Splone

Splone is a Berlin-based IT security research team and consultancy. We help improve IT-security by offering red team assements, penetration tests, audits and customized consulting.

Netragard

Netragard

Netragard has an established reputation for providing high-quality offensive and defensive security services.

Interos

Interos

Interos is the operational resilience company — reinventing how companies manage their supply chains and business relationships — through a breakthrough AI SaaS platform.

Viettel Cyber Security

Viettel Cyber Security

Viettel Cyber Security is an organization under the Military Telecommunication Industry Group, conducting research and developing information security solutions for domestic and foreign customers.

DartPoints

DartPoints

DartPoints helps bridge the digital divide by delivering cloud, colocation, managed services + edge infrastructure.

Detego Global

Detego Global

Detego Global are the creators of the Detego® Unified Digital Forensics Platform, a suite of modular tools used globally by military, law enforcement and intelligence agencies, and enterprises.

Quarkslab

Quarkslab

Quarkslab is a dedicated team of cyber-security engineers and developers. We aim at forcing the attackers, not the defender, to adapt constantly.

Cassini

Cassini

Cassini Cyber Threat Intelligence (CTI) helps protect your organisation from cyber attacks using threat intelligence from trusted New Zealand agencies.

Highway Ventures

Highway Ventures

HIGHWAY Ventures is a startup studio that builds cybersecurity and vertical AI companies in Northwest Arkansas from technology developed within the federal lab ecosystem.