State will not be able to implement regulations on artificial intelligence technologies for ten years under a plan under consideration in the U.S. House of Representatives. The law states in an amendment accepted this week by the House Energy and Commerce Committee that “we cannot enforce a law or regulation regulating an artificial intelligence model, an artificial intelligence system, or an automated decision system for ten years.” The proposal would still require approval from both Congressional Rooms and President Donald Trump before it becomes law.
AI developers and some lawmakers say federal action is needed to prevent the nation from creating a patchwork of various rules and regulations that could slow the growth of technology. Since CHATGPT exploded in the scene in late 2022, the rapid growth of generated AI has led companies to adapt their technology to as many spaces as possible. Economic implications are important as the US and China compete to see which countries’ technology is dominant, but they pose privacy, transparency and other risks to consumers that lawmakers have tried to ease.
Alexandre Wang, founder and CEO of data company Scale AI, told lawmakers at an April hearing. “But we need to be clear about one federal standard and there’s a preemptive measure to prevent this outcome, with 50 different standards.”
Efforts to limit states’ ability to regulate artificial intelligence could mean less consumer protection for technologies that are increasingly permeating all aspects of American life. “There was a lot of debate at the state level, and I think it’s important to approach this issue at multiple levels,” said Anjana Susarla, a professor at Michigan State University who studies AI. “You can approach it at the national level. You can approach it at the state level too. I think you need both.”
Several states have already begun adjusting AI
The proposed language prohibits enforcement of regulations, including those already listed in the book. The exceptions are rules and laws that facilitate AI development, and rules and laws that apply the same standards to non-AAI models and systems that do the same thing. These types of regulations are already beginning to pop up. The biggest focus is not in the US, but Europe, where the European Union already implements AI standards. However, the nation is beginning to get caught up in action.
Colorado passed a series of consumer protections last year, and is expected to come into effect in 2026. Last year, California adopted more than 12 AI-related laws. Other states have laws and regulations that frequently deal with certain issues such as deepfakes and the release of information about AI developers about training data. At the local level, some regulations also address potential employment discrimination when AI systems are used for employment.
“When it comes to what you want to regulate with AI, the state is on the map,” said Arsen Kourinian, a partner at law firm Mayer Brown. So far, state legislators have introduced at least 550 proposals on AI, according to the national meeting of the state legislature. At a House Committee hearing last month, Rep. Jay Obernolte, a Republican from California, expressed his desire to go beyond more state-level regulations. “We have a limited legislative runway to ensure that the state can resolve that issue before it goes too far,” he said.
Some states have book laws, but not all of them have come into effect or seen enforcement. This limits the potential short-term impact of the moratorium, says Covensway Felkeegan of the International Association of Privacy Professionals, Washington’s Managing Director. “There is no execution yet.”
The moratorium is likely to block state lawmakers and policymakers from developing and proposing new regulations, Zweifel Keegan said. “The federal government will be the only major potential regulator around the AI system,” he said.
What does a suspension on state AI regulations mean?
AI developers are demanding guardrails placed in their work to be consistent and streamlined. During a Senate Commerce Committee hearing last week, Openai CEO Sam Altman told Sen. Ted Cruz, a Republican from Texas, that the EU-style regulatory system is “disastrous” for the industry. Instead, Altman suggested that the industry develop its own standards.
From Sen. Brian Schatz, a Hawaii Democrat, Altman thought some guardrails would be good if the industry’s self-regulation is currently sufficient, but said, “It’s easy to go too far. Just as I’ve learned more about how the world works, I’m afraid it could go too far and have really bad results.” (Disclosure: CNET’s parent company Ziff Davis filed a lawsuit against Openai in April, claiming it infringed Ziff Davis’ copyright in training and operating AI systems.)
Concerns from businesses – both developers who create AI systems and “deployers” who use them in their interactions with consumers – are often attributed to fears that states will require important tasks such as impact ratings and transparency notices before the product is released. Consumer advocates say more regulations are needed, and hampering state capabilities could undermine user privacy and safety.
“AI is widely used to make decisions about transparency, accountability, or the lives of people who are not relying on, and also promotes fraud, impersonation and surveillance,” said Ben Winters, director of AI and Privacy at the American Consumer Federation. “A 10-year suspension leads to more discrimination, more deceptions and less control. Simply put, it’s with tech companies more than the people they clash with.”
Kourinian said suspensions on certain state rules and laws could result in more consumer protection issues being addressed by courts or the state attorney general. Existing laws regarding unfair and deceptive practices inherent in AI still apply. “Time tells us how judges interpret these issues,” he said.
Suzara said the spread of AI across the industry means that states could potentially be able to regulate issues such as privacy and transparency more widely without focusing on technology. However, suspensions on AI regulations could lead to such policies being bound by lawsuits. “We need to find some kind of balance between ‘I don’t want to stop innovation’, but we need to realize that there may be real consequences,” she said.
According to Zweifel-Keegan, many policies regarding the governance of AI systems occur due to these so-called technology-independent rules and laws. “It’s worth remembering that there are a lot of existing laws, and it could create new laws that don’t cause moratoriums, but apply to AI systems as long as they apply to other systems,” he said.
