OpenAI, Google và Meta đang vận động Nhà Trắng và Quốc hội Mỹ nhằm ngăn chặn các bang ban hành luật AI

• Các tập đoàn công nghệ lớn như OpenAI, Google, Meta đang chuẩn bị chiến dịch ngăn cản Mỹ ban hành luật AI ở cấp bang.

• Động thái này diễn ra sau khi 5 bang đã thông qua luật AI quan trọng:

  • Colorado: luật nghiêm ngặt nhất, yêu cầu tài liệu, kiểm định đảm bảo không phân biệt đối xử. Tuy nhiên, Thống đốc Jared Polis triệu tập phiên đặc biệt xem xét sửa đổi hoặc hủy bỏ vì lo ngại cản trở đổi mới.

  • California: yêu cầu công khai dữ liệu huấn luyện và gắn nhãn nội dung do AI tạo ra.

  • Texas: cấm AI phục vụ thao túng hành vi, phân biệt đối xử, tạo nội dung khiêu dâm trẻ em.

  • Tennessee: thông qua “Elvis Act” – cấm bắt chước giọng nói người khác bằng AI nếu không có phép.

  • Utah: buộc công bố khi cá nhân đang tương tác với AI tạo sinh “nguy cơ cao”.

• Ngoài ra, New York vừa thông qua dự luật an toàn công cộng, buộc Big Tech giảm nguy cơ sản phẩm AI gây “tổn hại nghiêm trọng”, đang chờ Thống đốc Hochul ký.

• Hiện có khoảng 500 dự luật AI đang được xem xét khắp nước Mỹ.

• Ở cấp liên bang, các nỗ lực ban hành luật AI toàn diện đã thất bại do chia rẽ chính trị. Đảng Cộng hòa từng cố gắng gắn lệnh cấm 10 năm cho luật bang vào dự luật thuế của Trump nhưng không thành công.

• Tháng 7/2025, Nhà Trắng đưa ra kế hoạch AI của Trump, trong đó khuyến nghị ngừng cấp vốn liên bang cho các bang có luật AI “quá hạn chế”.

• Lập luận từ Big Tech và quỹ VC (Andreessen Horowitz): nên tập trung quy định cách sử dụng AI thay vì chặn phát triển, vì ngành này có thể đem lại nghìn tỷ USD vốn hóa.

• Các luật sư cảnh báo nếu mỗi bang có một bộ quy định riêng sẽ tạo thành “miếng vá luật” (patchwork regulation), khiến doanh nghiệp khó mở rộng và luật tụt hậu so với tốc độ đổi mới.

• Một số nhà lập pháp bang phản ứng gay gắt: “Chính quyền liên bang không thể ngăn chúng tôi bảo vệ công dân,” phát biểu của Brandon Guffey (South Carolina).

📌 Mỹ đang đối diện cuộc đối đầu pháp lý giữa Big Tech và các bang: trong khi 5 bang đã có luật AI và hơn 500 dự luật khác đang chờ, các tập đoàn như OpenAI, Google, Meta muốn ngăn chặn luật AI riêng tại 50 bang và chuyển trọng tâm sang quy định cách dùng AI. Tháng 7/2025, Nhà Trắng đưa ra kế hoạch AI của Trump, trong đó khuyến nghị ngừng cấp vốn liên bang cho các bang có luật AI “quá hạn chế”. Các bang phản kháng để giữ quyền bảo vệ người dân. Cuộc chiến luật AI Mỹ dự báo sẽ ngày càng căng thẳng.

https://www.bloomberg.com/news/articles/2025-08-21/tech-industry-girds-for-new-battle-to-block-state-ai-rules

Tech Giants Prepare to Challenge State AI Regulations

By Emily Birnbaum
August 21, 2025 at 12:00 PM UTC
Tech companies including OpenAI, Meta Platforms Inc. and Alphabet Inc.’s Google are revving up to block US states from regulating their fast-moving and lucrative artificial intelligence businesses.
The tech giants are pushing both the AI-friendly White House and Republican-led Congress after five states, including Texas and California, passed significant AI-related laws.
“Lawmakers should be inviting innovation, not driving it away from the state,” said Kouri Marshall, director of state government relations with tech trade group Chamber of Progress, which counts Andreessen Horowitz, Google, Apple Inc. and Amazon.com Inc. among its members.
While few of the farthest-reaching state rules have gone into full effect, advocates for the industry want to redirect the regulation to the ways AI is used and away from the development of new AI. The first companies to make AI breakthroughs potentially stand to gain trillions of dollars in market capitalization.
“My hope is the focus moves away from trying to regulate development to regulating use” by individual customers, said Matt Perault, head of artificial intelligence policy at venture capital firm Andreessen Horowitz.
Republicans in Congress tried and failed in June to attach a 10-year ban on enforcing state regulations onto President Donald Trump’s tax legislation.

Trump Wins Broad Economic Policy Shift As House Passes Tax Bill
President Donald Trump displays his centerpiece tax legislation after signing it during a ceremony at the White House on July 4.Photographer: Bonnie Cash/UPI/Bloomberg
Undeterred, the companies successfully pushed White House advisers to include a version of the federal moratorium on state regulation in Trump’s AI plan in July. The plan gives non-binding guidance to federal agencies, including a directive that AI-related federal funding shouldn’t go to states with “unduly restrictive” AI regulation.
Tech company representatives also are trying again to attach the 10-year ban to future legislation, said industry lobbyists who spoke on condition of anonymity.
As AI use grows, taking on roles such as assessing job applications, identifying criminal suspects, handling medical claims and creating images nearly impossible to distinguish from genuine photos or video, state lawmakers are eager to impose some rules of the road. Federal legislation has largely stalled amid partisan disagreements and pushback from tech-friendly lawmakers.
Read more: Why Are Deepfakes Everywhere? Can They Be Stopped?
Some states are considering legislation that would require companies to conduct audits ensuring their systems don’t harm consumers, disclose when people are interacting with AI and bar the companies from copying artists’ creative work.
But tech companies and venture capital firms backing AI startups fear any regulation could slow their growth in a new sector. And if they have to be regulated, they’d prefer to avoid navigating 50 different state rules.
Hope Anderson, a privacy and AI lawyer with White & Case, said the speed of technological change makes it “tricky” for the law to keep up, and more so if states enact a “patchwork” of differing regulations.
State AI Laws
Some 500 laws are under consideration around the US that would affect private companies developing and deploying AI, according to Goli Mahdavi, an AI lawyer in San Francisco at Bryan Cave Leighton Paisner LLP.

Samsung Nears Wide-Ranging Deal With Perplexity For AI Features
As AI spreads through society, state lawmakers are increasingly anxious to impose some rules of the road.Photographer: Stefani Reynolds/Bloomberg
Yet only five states have laws on the books that significantly impact how tech companies do business. Other states have passed more limited legislation regulating how businesses can use AI for specific purposes such as employers in screening job applications or health providers in making medical diagnoses.

Get the Tech Newsletter bundle.
Bloomberg's subscriber-only tech newsletters, and full access to all the articles they feature.

 Subscribe
Colorado’s AI law goes the farthest, requiring developers and companies to provide extensive documentation and conduct tests to ensure their systems don’t discriminate against users based on protected characteristics like race or gender. Yet tech industry lobbyists persuaded Governor Jared Polis and state lawmakers to reopen the law to amendments during a special session scheduled to begin Thursday. Polis said in a statement he’s concerned about the law’s impact on “critical technological advancements.” Lawmakers are considering a total repeal of the law or reducing the number of AI systems it applies to.
California’s narrower law requires AI developers to publicly post information about the data used to train AI systems and notify users when posts, images or videos are AI-generated.
Texas’s law restricts the development and deployment of AI systems for behavioral manipulation, discrimination, or the creation and distribution of child pornography.
Tennessee’s “Elvis Act” prohibits the use of AI to mimic a person’s voice without their permission, while Utah requires developers of “high-risk” AI systems to disclose that individuals are interacting with generative AI, not a human.
“Most of the concern flows from the potential for more laws” rather than existing ones, said Cobun Zweifel-Keegan, managing director of the International Association of Privacy Professionals.
States to Watch
The New York legislature this summer passed a public safety bill that would require the largest tech companies to reduce risk of their products causing “critical harm.” It’s now with Governor Kathy Hochul, who has not said whether she plans to sign the legislation into law.California Governor Gavin Newsom in 2024 vetoed the most comprehensive effort yet to regulate AI, which would have compelled companies to test whether their AI models would lead to mass death, endanger public infrastructure or enable cyberattacks. California lawmakers are considering narrower measures, including a requirement companies notify individuals when automated systems make decisions about them and more oversight of “high-risk” automated systems.
Filling Vacuum
The states are acting to fill the vacuum left by the federal government. “There’s a desire for states to regulate in this space as it’s increasingly clear that there will not be a federal omnibus AI law,” said Mahdavi.
And in some cases, the industry’s drive to use Congress to override state laws is stirring local lawmakers to move more quickly.
“I’ll be damned if the federal government is going to prevent me from protecting citizens within my state,” said South Carolina Representative Brandon Guffey, who chairs a state House subcommittee on AI. “People are scared they’re going to put in this moratorium so they’re trying to get bills passed before they do.”