Big Tech's Last Stand on AI Rules

The White House is reportedly pushing for a swift, politically expedient alliance between child safety advocates and proponents of artificial intelligence regulation, as the midterm elections draw near.

For an extended period, Washington lobbyists representing major technology companies have pursued a primary legislative objective: federal preemption for AI. This sought-after outcome involves a unified federal law, enacted by Congress and signed by the President, establishing a single framework for AI regulation nationwide, thereby superseding the complex and fragmented state-level regulatory landscape. However, these lobbying efforts have encountered significant resistance and widespread political criticism, leading to concerns that a potential post-midterm shift in congressional control to less amenable Democrats could further hinder their progress.

Now, their most critical and perhaps last-ditch attempt at achieving preemption is being complicated by an unrelated, pre-existing legislative battle in Congress concerning child online safety, a debate that began before the public release of ChatGPT.

Reports surfaced earlier this week indicating that the White House communicated to both child safety organizations and major tech firms its intention to support a collection of children's online safety legislation championed by Senator Marsha Blackburn (R-TN), co-author of the Kids Online Safety Act (KOSA), as an integral component of a broader AI preemption initiative. While online safety shares relevant intersections with AI, it represents just one dimension of a much more extensive and intricate array of concerns—including frontier model safety, discrimination, and environmental impact—that a truly comprehensive AI law would need to encompass.

Nevertheless, this prospective agreement has encountered a significant obstacle: the White House seemingly failed to notify House Republicans, who had recently passed their own iteration of KOSA, that it intended to utilize Senator Blackburn’s legislation as the primary vehicle. Democrats who had collaborated with Blackburn on the Senate’s version of KOSA were reportedly also kept uninformed. Furthermore, a distinct, bipartisan-supported AI preemption bill was already under consideration in the House.

This situation led to a week of profound uncertainty for advocates of both policy objectives. While the bundling of AI preemption and child safety might be a strategy to ensure preemption's enactment, the specific version of child safety legislation to be adopted remains ambiguous. Questions arose: Would it be the more stringent Senate KOSA, or the less restrictive version supported by House Majority Leader Steve Scalise (R-LA)? Moreover, the White House's precise role in these unfolding developments was unclear.

"No one truly understands who is orchestrating this effort," a Republican lobbyist representing a midsize technology firm conveyed to The Verge. "There is profound skepticism regarding the bill's advancement, given the disparate viewpoints among stakeholders. I believe the House will not proceed with any legislation favored by Blackburn."

Despite the significant divisions within GOP leadership and its populist faction over AI regulation, former President Donald Trump has publicly advocated for the enactment of an AI preemption bill, underscoring the Republican Party's imperative to achieve this outcome. Currently, White House policy strategists are endeavoring to refine a preemption strategy, drawing influence from Mike Davis, a lawyer allied with Trump and founder of the Article III Project, who notably led a successful effort to defeat a different AI moratorium in the Senate the previous year.

Generally, to secure Davis’s endorsement, any preemption legislation must substantively safeguard a set of principles he terms the "Four Cs": children, conservatives, creators, and communities. While some of these tenets were incorporated into the White House’s comprehensive AI law draft, unveiled in March, and KOSA's inclusion addressed the "children" criterion, Davis explicitly informed The Verge that he expects any proposed legislation to encompass all four. He stated, "There is no chance in hell AI preemption will pass if it does not address the Four Cs. I will make damn sure of that. Again."

However, advancing KOSA necessitates resolving significant disparities between the House and Senate iterations of the bill. The Senate’s version mandates that technology companies undertake a "duty of care," implementing proactive measures to safeguard young users, and extends this obligation to AI enterprises. Conversely, the House version, primarily driven by Scalise, significantly weakened this provision in late November, provoking outrage among child safety advocates. Consequently, the House’s apparent exclusion from the White House’s deliberations struck observers as particularly noteworthy. Michael Toscano, senior fellow and director of the Family First Technology Initiative for the conservative Institute for Family Studies, observed, "[Blackburn] genuinely does not want House KOSA."

Even if former President Trump could sway House Republicans, another hurdle remains: congressional Democrats. They, too, became aware of Blackburn’s negotiations with the White House concurrently with their House Republican counterparts. While Senate KOSA, co-sponsored by Senator Richard Blumenthal (D-CT), passed overwhelmingly 91–3 in 2024, Democrats were reportedly unaware that their legislation would now be linked to the contentious objective of AI preemption. An AI policy advocate remarked, "If they [Blackburn and the White House] are looking at a standalone bill, it’ll have to go through the Senate," emphasizing that any new iteration of such a bill would necessitate 60 votes, thus requiring Democratic support for passage.

Furthermore, even if the bill were to garner some support, the legislative calendar presents a significant challenge. "It is mid-June. You have a month and a half before people leave for [five-week] recess. And then it’s [general] election season," the AI policy advocate explained, concluding, "There’s just no way." The remaining legislative time is already consumed by pressing issues, including the renewal of FISA, an immigration enforcement package, increased defense allocations for Trump’s proposed conflict with Iran, a cryptocurrency market structure bill, affordability initiatives, the contentious SAVE America election bill, and essential budgetary items such as Medicaid.

The intertwining of AI preemption and KOSA presents a complex dilemma for Big Tech: Is a comprehensive federal AI preemption more desirable than protection from a "duty of care" obligation? The Republican tech lobbyist highlighted the limited timeframe for this decision, particularly if Democrats gain control of a chamber. "After the election, what incentive do the Democrats have to support anything? Why wouldn’t they simply assert, 'We’re going to pursue our own agenda in the new Congress?' I am profoundly skeptical."

Austin Carson, previously the head of government relations at Nvidia and founder of SeedAI, a nonprofit dedicated to expanding AI access for local communities, expressed greater doubt regarding the success of this opportunistic KOSA-preemption alliance. He conveyed to The Verge, "I can’t imagine a scenario where [this bill] would move. I cannot imagine it."