David Sacks: The AI Regulatory Frenzy at the State Level is “Very Concerning”

A rapid increase in state-level AI regulations, like Colorado's SB 24-205, is creating a complex compliance burden for startups and could lead to 'woke AI' by forcing models to prevent disparate impact.

• 50 states are rapidly introducing and passing diverse AI regulations without clear consensus.

• Colorado's new law, SB 24-205, bans 'algorithmic discrimination' with broad liability.

• Compliance with these laws may force AI models to prevent outputs with disparate impact, potentially leading to 'woke AI'.

A 'regulatory frenzy' across all 50 states is introducing thousands of AI bills, with 118 laws already passed, driven by an imperative to act without clear consensus on risks or solutions. This creates a challenging compliance environment, particularly for startups facing 50 different reporting regimes. Colorado's new law, SB 24-205, exemplifies this trend by banning 'algorithmic discrimination' and imposing liability on both AI developers and deployers, potentially forcing models to alter truthful outputs to avoid disparate impact.

State Regulatory Frenzy

• 00:00:00 A concerning 'regulatory frenzy' is unfolding at the state level, with all 50 states introducing AI bills in 2025, totaling over a thousand proposals and 118 already enacted. This rush to regulate AI lacks clear agreement on risks or objectives, leading to a fragmented landscape of 50 different reporting regimes that will trap startups trying to comply. The sheer volume and inconsistency of these laws create significant compliance challenges.

Colorado's AI Law

• 00:00:46 Colorado's SB 24-205, the Consumer Protections for Artificial Intelligence Act, bans 'algorithmic discrimination,' defined as unlawful differential treatment or disparate impact based on protected characteristics like age or race. This law holds both AI model developers and deployers liable if an AI decision, even one based on race-neutral criteria, results in a disparate impact on a protected group. Compliance might force developers to integrate a DEI layer into models, preventing truthful outputs that could be perceived as discriminatory, leading to 'woke AI'.