US president Donald Trump's "One Big Beautiful Bill", passed by the House of Representatives on Thursday, brought in a whole host of changes, from tax breaks and immigration crackdowns to tweaking welfare programmes. But it also had a significant AI-related provision snuck in - a 10-year moratorium on AI-related state legislations.
If the bill is now passed by the Senate and approved by the president, it would enact one of the most consequential federal tech policy actions in years.
The provision has sparked a debate between those opposing a "patchwork" of state laws and those calling to address harms in the absence of a federal AI law.
The bill talks about how the commerce department may use funds to leverage AI for operational efficiency and cybersecurity. It then adds that no state or political subdivision may enforce any regulations on AI models or automated decision systems for a 10-year period after enactment of the bill.
This does not apply to regulations that are aimed at removing legal impediments to AI deployment and enabling adoption and procurement of the technology, it says. It also does not curb regulations that don't impose any substantive design, performance, data-handling, documentation, civil liabilities, taxation or other requirements on AI systems.
The moratorium is seen as benefiting the technology industry, many of whom have called for a regulatory environment that fosters freedom to innovate.
The industry appreciates the government's "recognition that a patchwork of state laws on AI could jeopardise national AI priorities," said Jason Oxman, president and CEO of tech industry body ITI Council, in a letter to a subcommittee involved in the legislation on Wednesday. "The technology industry is increasingly concerned about the growing number of state legislative proposals to regulate AI," Oxman wrote, adding that competing state regulations could hamper efficiency gains possible from greater government adoption of AI. "There is a risk that a tidal wave of state legislative activity could undermine our shared goals of American technological preeminence."
ITI, which counts Microsoft, Google, IBM, Accenture, Adobe, Meta and OpenAI among its members, advocated for a unified, risk-based standard for AI development and deployment that avoids conflicting state-level regulation.
Tech giant Google in a March submission regarding the US government's "AI Action Plan" which aims to strengthen US leadership in AI, said that it should "preempt a chaotic patchwork of state-level rules on frontier AI development."
ChatGPT-maker OpenAI proposed a "holistic approach" that enables voluntary partnership between the federal government and the private sector, without American AI companies having to comply with "overly burdensome state laws".
Analysts say that the bill reflects the administration's deregulatory approach.
On the other hand, in the run up to this bill being introduced, a group of 40 attorney generals of states like California, Colorado, Washington, New Jersey, and others wrote to US Congress officials opposing the moratorium.
"Such a broad moratorium would be sweeping and wholly destructive of reasonable state efforts to prevent known harms associated with AI," they said.
In contrast to the EU, which has an overarching AI law, the US which has no federal AI law was seen as having a "light-touch" approach.
But in reality, AI-related bills in US states have been ballooning. A report by tech industry body BSA, the Software Alliance, showed that in 2024, nearly 700 AI-related bills were introduced across 45 US states. About 113 have been enacted. In 2023, there were under 200 such bills in states. The momentum was expected to carry into 2025 as well.
Colorado was the first to pass comprehensive legislation to address high-risk uses of AI. States like California and Tennessee have enacted targeted reforms to address specific issues such as data provenance and digital replicas.
Forty-three states considered legislation to address issues around AI and deepfakes. Of over 300 such pieces of legislation introduced, 71 were enacted into law.
In India, neither central nor state governments have introduced AI legislation. The government expects existing laws and sectoral regulations to be sufficient in tackling harms. Officials have said that AI legislation may be introduced in future if the need arises.
If the bill is now passed by the Senate and approved by the president, it would enact one of the most consequential federal tech policy actions in years.
The provision has sparked a debate between those opposing a "patchwork" of state laws and those calling to address harms in the absence of a federal AI law.
The bill talks about how the commerce department may use funds to leverage AI for operational efficiency and cybersecurity. It then adds that no state or political subdivision may enforce any regulations on AI models or automated decision systems for a 10-year period after enactment of the bill.
This does not apply to regulations that are aimed at removing legal impediments to AI deployment and enabling adoption and procurement of the technology, it says. It also does not curb regulations that don't impose any substantive design, performance, data-handling, documentation, civil liabilities, taxation or other requirements on AI systems.
The moratorium is seen as benefiting the technology industry, many of whom have called for a regulatory environment that fosters freedom to innovate.
The industry appreciates the government's "recognition that a patchwork of state laws on AI could jeopardise national AI priorities," said Jason Oxman, president and CEO of tech industry body ITI Council, in a letter to a subcommittee involved in the legislation on Wednesday. "The technology industry is increasingly concerned about the growing number of state legislative proposals to regulate AI," Oxman wrote, adding that competing state regulations could hamper efficiency gains possible from greater government adoption of AI. "There is a risk that a tidal wave of state legislative activity could undermine our shared goals of American technological preeminence."
ITI, which counts Microsoft, Google, IBM, Accenture, Adobe, Meta and OpenAI among its members, advocated for a unified, risk-based standard for AI development and deployment that avoids conflicting state-level regulation.
Tech giant Google in a March submission regarding the US government's "AI Action Plan" which aims to strengthen US leadership in AI, said that it should "preempt a chaotic patchwork of state-level rules on frontier AI development."
ChatGPT-maker OpenAI proposed a "holistic approach" that enables voluntary partnership between the federal government and the private sector, without American AI companies having to comply with "overly burdensome state laws".
Analysts say that the bill reflects the administration's deregulatory approach.
On the other hand, in the run up to this bill being introduced, a group of 40 attorney generals of states like California, Colorado, Washington, New Jersey, and others wrote to US Congress officials opposing the moratorium.
"Such a broad moratorium would be sweeping and wholly destructive of reasonable state efforts to prevent known harms associated with AI," they said.
In contrast to the EU, which has an overarching AI law, the US which has no federal AI law was seen as having a "light-touch" approach.
But in reality, AI-related bills in US states have been ballooning. A report by tech industry body BSA, the Software Alliance, showed that in 2024, nearly 700 AI-related bills were introduced across 45 US states. About 113 have been enacted. In 2023, there were under 200 such bills in states. The momentum was expected to carry into 2025 as well.
Colorado was the first to pass comprehensive legislation to address high-risk uses of AI. States like California and Tennessee have enacted targeted reforms to address specific issues such as data provenance and digital replicas.
Forty-three states considered legislation to address issues around AI and deepfakes. Of over 300 such pieces of legislation introduced, 71 were enacted into law.
In India, neither central nor state governments have introduced AI legislation. The government expects existing laws and sectoral regulations to be sufficient in tackling harms. Officials have said that AI legislation may be introduced in future if the need arises.
