End of Self-Regulation? U.S. Senate Pushes for 'Duty of Care' for AI Companies

The U.S. Congress is moving to impose a 'duty of care' on advanced AI developers to prevent large-scale harm.
The U.S. government will be granted the authority to block the release of risky models, with participation from national research institutes for safety verification.
The regulatory costs could burden release and investment schedules, potentially favoring large AI companies.

[Block Media Reporter Lee Hye-yeon] The U.S. Congress is discussing a plan to convert the safety management of cutting-edge artificial intelligence (AI) developers from a voluntary commitment to a legal obligation. The possibility of the federal government blocking the release of AI models that pose a risk of causing large-scale harm is also under consideration.

According to Reuters, the U.S. Senate negotiating team is discussing a bill that would impose legal responsibilities on developers of cutting-edge AI models to prevent known significant risks.

If the bill passes, significant changes are expected in the safety management standards of major AI companies such as Google, Anthropic, and OpenAI. While there are expectations that stricter regulations could enhance consumer and investor confidence, there are also concerns that only a few large companies capable of bearing the enormous regulatory costs may strengthen their market dominance.

The Core of the Bill: Imposing 'Duty of Care' on AI Developers

The core of the proposed bill is to impose a so-called 'duty of care' on frontier AI model developers. Frontier AI refers to the most advanced AI models among existing technologies.

The bill includes provisions for designing AI models to minimize the risk of large-scale harm from the development stage. It also includes the U.S. government's authority to prevent the release of unsafe AI models.

A procedure will be established for companies to challenge government decisions in court. There are also discussions about involving experts from national research institutes and government agencies in the verification and testing of AI models.

This federal bill may take precedence over some state laws that regulate similar issues. The negotiations involve Senate Republican Leader John Thune, Senate Commerce Committee Chairman Ted Cruz, Senator Amy Klobuchar, and Senator Maria Cantwell.

Senator Klobuchar stated to Reuters, "The goal is for the government to oversee the most serious risks that AI models can pose." She explained that developers should collaborate with government experts when validating and testing advanced models.

Senator Cruz also made a similar proposal. He stated via social media X that he, along with Senator Klobuchar and Leader Thune, is addressing catastrophic risks related to biological or nuclear threats.

AI is advancing at an extraordinary pace, and some of the risks are dangerous and frightening.

I've been on the forefront of this issue in the Senate. I wrote and passed the TAKE IT DOWN Act, the only legislation Congress has enacted restricting AI, to protect victims from... pic.twitter.com/movYZNQrHr

--- Senator Ted Cruz (@SenTedCruz) September 9, 2026

Whether federal regulations will replace some state regulations is a major issue. This aligns with the Trump administration's legislative direction on AI, which has called for a consistent national regulatory framework, arguing that differing regulations across states could hinder the development of the U.S. AI industry.

The Reality of AI Risks... Europe is Already Regulating

U.S. lawmakers are not discussing regulations based solely on hypothetical scenarios. Instances where AI agents have accessed external systems on their own and concerns raised by major AI company researchers have influenced legislative discussions.

The International Association of Privacy Professionals (IAPP) reported that the resignation of former Anthropic researcher Jacob Coxon and his concerns about the competitive development of self-improving AI sparked negotiations.

Regulatory demands within the U.S. Congress are also increasing. Senator Josh Hawley is investigating OpenAI's role in the cyberattack on Hugging Face that occurred in July. Senator Bernie Sanders is preparing a bill to temporarily halt the development of superintelligent AI.

OpenAI also supports mandatory federal regulations. Chris Lehane, OpenAI's global policy chief, stated in a policy document released on the 9th that "mandatory national safety regulations applicable according to the level of AI capabilities" are necessary. This indicates that both some in the AI industry and Congress share concerns about whether voluntary safety measures are sufficient.

The European Union (EU) has already implemented regulations for high-risk AI models. The EU AI Act imposes obligations for model evaluation, risk management, incident reporting, and cybersecurity on providers of general-purpose AI models that could pose systemic risks.

Regulatory Costs Impacting Capital Markets... Favoring Large AI Companies

The impact of AI regulations on capital markets is also significant. Goldman Sachs Research predicts that global AI investment will reach approximately $1 trillion by 2026, with U.S. investments projected at $581 billion.

The introduction of new testing, legal reviews, and documentation obligations could affect AI model release schedules, infrastructure budgets, corporate valuations, and IPO timings.

For small and medium-sized enterprises that may struggle to bear compliance costs, this could act as a barrier to entry. Ultimately, this means that large AI companies with ample financial resources may become more advantageous.

AI safety issues have already emerged as a topic in capital markets. According to Cryptopolitan, David Sacks has argued that Anthropic should halt its IPO until the safety-related allegations raised by former researcher Coxon are confirmed.

Anthropic investors are reportedly discussing a corporate valuation of about $2 trillion. Consequently, the safety of frontier AI has become a significant issue not only for regulators but also for investors.

However, it remains uncertain whether the bill will actually pass through Congress. According to Reuters, the U.S. House of Representatives plans to hold only one week of sessions before the midterm elections on November 3, while the Senate is expected to continue its schedule for three weeks.

Even if the Senate negotiating team reaches an agreement, the limited remaining session schedule could pose a major obstacle to the bill's passage.

This content is provided for general informational purposes only and doesn't constitute financial, investment, legal, or tax advice. Any events, rewards, online promotions, or related information mentioned herein should not be considered a recommendation, solicitation, or invitation to purchase, sell, trade, or otherwise deal in any crypto assets. Crypto assets are highly volatile and may result in loss. The availability of WEEX services, products, and related events may vary by region. You are responsible for ensuring that your participation is in accordance with applicable local laws and regulations.