Meta Spent $6.5M to Kill Mandatory AI Safety Audits

In the final three months of 2025, Meta Platforms Inc. spent a record $6.5 million on federal lobbying, a blitz that effectively dismantled the primary enforcement mechanism of the 2026 AI Oversight Act. This spending surge, documented in late-year LD-2 disclosure filings, targeted a specific legislative pivot: the transformation of 'Mandatory Algorithmic Accountability' into a series of 'Voluntary Safety Frameworks.' While mainstream outlets have framed Meta’s involvement as 'proactive engagement' with regulators, the paper trail reveals a targeted financial campaign to ensure the emerging AI industry remains self-regulated.

According to January 2026 Federal Election Commission (FEC) filings, the influence campaign relied heavily on [Lobbyist Bundling], which is the practice of professional lobbyists collecting individual contributions from various donors and delivering them in a single package to a candidate's campaign. Meta’s strategy exploited a new 2026 FEC disclosure threshold. Under the updated rules for Form 3L, lobbyists must disclose when they provide bundled contributions exceeding $24,000. Within a three-week window in January, 14 members of the House Energy & Commerce Committee—the very body presiding over the AI Oversight Act—received bundled packages from Meta-linked lobbyists totaling $336,000.

This influx of capital preceded a closed-door markup session where Section 402 of the Act was fundamentally rewritten. Originally, the section mandated independent audits of AI models to prevent bias in lending, housing, and hiring. Following Meta’s record spend, the word 'mandatory' was struck from the text. The new language allows corporations to conduct their own 'internal safety assessments' without federal verification. [Algorithmic Accountability] is the legal requirement for companies to audit and explain the decision-making processes of their AI systems to ensure they do not discriminate or cause harm. By shifting to a voluntary model, the legislation removes the threat of legal penalties for firms that deploy biased or flawed technology.

Meta’s influence is reinforced by a robust [Revolving Door] strategy, a term describing the movement of high-level employees between public service and private sector lobbying. Of the 42 lobbyists Meta employed in late 2025, 65% previously held staff or leadership positions within the Senate Judiciary or House Energy & Commerce committees. These lobbyists, directed by Meta’s VP of Global Public Policy Joel Kaplan, leveraged their internal connections to push for the removal of the 'right to explanation' clause. Meta’s own LD-2 disclosure documents explicitly requested this removal, citing 'trade secret protection' as the justification for keeping AI decision-making processes opaque to the public.

The $6.5 million quarterly expenditure is the highest in the company’s history, surpassing previous peaks during the 2021 antitrust hearings. The money was funneled through a network of K Street firms that specialize in navigating the House Energy & Commerce Committee’s leadership. Records show that of the $336,000 bundled to committee members, over 70% went to representatives who subsequently voted to approve the softened 'Voluntary Safety Frameworks' during the final committee vote. This correlation suggests a high return on investment for Meta’s lobbying arm.

What the mainstream press has omitted is the specific nature of the 'right to explanation' clause that Meta successfully lobbied against. This clause would have required AI-driven platforms to provide a plain-language reason to any user denied a loan, a job interview, or an apartment due to an automated decision. Without this right, ordinary citizens are left with no legal recourse when a 'black box' algorithm makes a life-altering error. Public safety and civil rights are essentially being outsourced to the companies that stand to profit most from their deregulation.

For the average person, this legislative shift means that the automated systems increasingly managing their lives will remain beyond the reach of the law. If an AI system incorrectly flags you as a credit risk or filters your resume out of a job pool, you no longer have a federal right to know why or to challenge the data used to make that choice. The 'Voluntary Safety Frameworks' championed by Meta ensure that corporate convenience takes precedence over individual transparency.

At Gen Us, we believe in following the money to see who is actually writing our laws. You can use our Politician Tracker to see if your representative was among the 14 committee members who accepted Meta-linked bundles. Our database includes full breakdowns of the January 2026 FEC Form 3L filings and a comparison of how these contributions aligned with specific votes on the AI Oversight Act.