Senator Markey proposes federal AI accountability bills

On July 11, 2026, Senator Ed Markey from Massachusetts introduced to Congress a set of federal bills aimed at regulating artificial intelligence. These laws are intended to restrict harmful applications of AI in four key areas, covering infrastructure, workplace, justice, and protection of vulnerable populations.

Watery data centers: first target

One of Markey's main focuses is the energy and environmental harm from AI data centers. Modern artificial intelligence models require enormous amounts of electricity and water for cooling servers. As demand for AI grows, data center resource consumption grows exponentially — some studies show that training a single large model may require millions of liters of water.

Markey's bill provides for strict transparency requirements: companies must publicly report water and energy consumption when developing and deploying AI systems. This will allow regulators to track the environmental impact of AI infrastructure locally and impose restrictions in water-scarce regions.

Workplace surveillance through AI

The second issue on Markey's agenda is the use of AI systems for mass monitoring of employees. Many companies already apply algorithms for automatic productivity tracking, emotion analysis through video cameras, predicting worker behavior, and identifying turnover risks.

This raises serious concerns about privacy, fairness, and psychological pressure on workers. Markey's bills will limit unjustified surveillance and require explicit informed consent from employees before implementing any AI monitoring systems.

Algorithms that discriminate

Markey's third vector is addressing AI bias in critical decisions affecting people's lives. Research shows that candidate search algorithms often discriminate by gender and race, filtering out qualified applicants from underrepresented groups. Credit scoring systems repeatedly deny minorities access to credit.

Even in criminal justice, AI risk assessment systems can amplify existing racial inequalities. The bills require independent audits of algorithms before use in hiring, lending, and court decisions, as well as appeal mechanisms for people affected by discriminatory decisions.

Protecting children from chatbots

The fourth target is psychological and commercial impact on minors. Chatbots, especially personalized AI-based systems, can manipulate children's emotions, creating dependency and collecting their behavioral data for targeted advertising.

Markey's law provides for bans on targeted targeting and data collection from children under 18 years old through AI systems trained to manipulate user behavior. This includes restrictions on recommendation algorithms that optimize engagement at the cost of youth psychological well-being.

Why federal level?

So far, AI regulation in the US has been going state by state, creating a patchwork of rules. California introduces one privacy standard, New York establishes different requirements for algorithm audits, other states do practically nothing.

For global technology companies, such administrative disorder forces them to apply the strictest requirements everywhere. Markey's federal approach will establish uniform rules for the entire country, simplifying compliance while ensuring a minimum level of citizen protection everywhere.

What this means

Markey's bills reflect Congress's growing concern about specific harmful applications of AI. If passed, they will become the first major federal measures aimed not at banning AI itself, but at limiting its worst applications and ensuring corporate accountability. This could trigger a wave of similar legislation in other countries, establishing new standards for AI governance and accountability.