\n<\/p>\n
The White House has announced a new AI policy framework that calls for Congress to craft federal regulation that overrules state AI laws. The Trump administration has made multiple attempts to overrule more restrictive state-level AI regulation, but has failed so far, most notably in the passing of the \u201cOne Big Beautiful Bill.\u201d<\/p>\n
The framework focuses on a variety of topics, covering everything from child privacy to the use of AI in the workforce. \u201cImportantly, this framework can succeed only if it is applied uniformly across the United States,\u201d The White House writes. \u201cA patchwork of conflicting state laws would undermine American innovation and our ability to lead in the global AI race.\u201d<\/p>\n
In terms of child privacy protections, the framework ask for Congress to require companies to provide tools like \u201cscreen time, content exposure and account controls\u201d while also affirming that \u201cexisting child privacy protections apply to AI systems,\u201d including limits on how data is collected and used for AI training. The framework also says carveout states should be allowed to enforce \u201ctheir own generally applicable laws protecting children, such as prohibitions on child sexual abuse material, even where such material is generated by AI.\u201d<\/p>\n
The energy-use and environmental impact of AI infrastructure is a going concern, but the White House\u2019s policy proposals are primarily worried about the cost of data centers. The framework suggests federal AI regulation should make sure that higher electricity costs aren\u2019t passed on to people living near data centers, while streamlining the process for permitting AI infrastructure construction, so companies can pursue \u201con-site and behind-the-meter power generation.\u201d The framework also calls for fewer restrictions on the software-side of AI development, proposing \u201cregulatory sandboxes for AI applications\u201d and asking Congress to \u201cprovide resources to make federal datasets accessible to industry and academia in AI-ready formats.\u201d<\/p>\n
While a recently AI bill from Senator Marsha Blackburn (R-Ten.) attempts to eliminate Section 230, a piece of a larger law that says platforms can\u2019t be held responsible for the speech they host, the framework appears to propose the opposite. \u201cCongress should prevent the United States government from coercing technology providers, including AI providers, to ban, compel or alter content based on partisan or ideological agendas,\u201d the White House writes. The framework is similarly hands-off when it comes to copyright and the use of intellectual property to train AI. \u201cAlthough the Administration believes that training of AI models on copyrighted material does not violate copyright laws,\u201d the White House writes, it supports the issue being settled in court rather than by legislation. Though, the White House does think Congress should \u201cconsider enabling licensing frameworks\u201d so IP holders can bargain for compensations from AI providers.<\/p>\n
The clincher in the White House\u2019s proposal is the idea that federal regulation should preempt state law, specifically so that states don\u2019t \u201cregulate AI development,\u201d don\u2019t \u201cunduly burden American\u2019s use of AI for activity that would be lawful if performed without AI\u201d and don\u2019t punish AI companies \u201cfor a third party\u2019s unlawful conduct involving their models.\u201d The idea that AI companies aren\u2019t liable for the illegal or harmful uses of their products is particularly problematic because it lies at the heart of multiple intersecting issues with AI right now, including it being used to generate sexually explicit images of children and allegedly playing a role in the suicide of users.<\/p>\n
Ultimately, though, the framework might be too contradictory to be useful, Samir Jain, the Vice President of Policy for the Center for Democracy and Technology, writes in a statement to Engadget:<\/p>\n
\n\nThe White House\u2019s high-level AI framework contains some sound statements of principles, but its usefulness to lawmakers is limited by its internal contradictions and failure to grapple with key tensions between various approaches to important topics like kids\u2019 online safety. It rightly says that the government should not coerce AI companies to ban or alter content based on \u2018partisan or ideological agendas,\u2019 yet the Administration\u2019s \u2018woke AI\u2019 Executive Order this summer does exactly that. On preemption, the framework asserts that states should not be permitted to regulate AI development, but at the same time rightly notes that federal law should not undermine states\u2019 traditional powers to enforce their own laws against AI developers. States are currently leading the fight to protect Americans from harms that AI systems can create, and Congress has twice correctly decided not to pursue broad preemption.<\/p>\n<\/div>\n<\/blockquote>\n
President Donald Trump has attempted to have an active role in how AI is developed and regulated in the US with mixed results, primarily because, as Jain notes, Congress has been unwilling to give up states\u2019 right to regulate the technology on their own terms. Without that, its hard to say how much of the framework will actually make it into federal law.<\/p>\n<\/div>\n