Congress Unveils Bipartisan Blueprint for Federal AI Oversight and Transparency

A bipartisan group of House lawmakers has released a discussion draft that could reshape how artificial intelligence is built and deployed in the United States. The proposed legislation, made public on June 4, 2026, aims to create a dedicated federal agency for AI governance while imposing new transparency and safety requirements on developers of large-scale models.

The bill targets companies building foundation models—the powerful systems behind everything from text generation to medical image analysis. Under the proposal, organizations whose models exceed certain computational thresholds would need to disclose details about training data, known biases, and safety testing before releasing products to the public.

A centerpiece of the draft is the creation of a new federal body with authority to set enforceable standards for high-risk AI applications, including those used in healthcare, lending, and critical infrastructure. Unlike earlier voluntary guidelines, this framework includes meaningful penalties for non-compliance.

Other notable provisions include mandatory incident reporting within 72 hours of a significant failure or security breach, explicit protections for intellectual property, and new privacy rules limiting the use of sensitive personal data in commercial AI training. The bill also funds retraining programs for workers affected by automation and directs the executive branch to strengthen export controls on advanced AI chips and technology.

Lawmakers borrowed elements from the European Union’s AI Act but tailored the approach to prioritize domestic innovation and protect trade secrets. The draft includes phased implementation timelines—basic transparency rules within one year, with more complex testing requirements rolling out over three to five years.

Industry groups responded cautiously, acknowledging the need for regulation while warning against overly burdensome rules that could slow development. Civil society organizations welcomed the accountability measures but pushed for stronger protections against bias and privacy violations.

This discussion draft is only the starting point for what promises to be a lengthy legislative process. The House committee plans additional hearings and stakeholder sessions before introducing a formal bill.