Senator Markey's AI Bills Target Hiring Bots, Water Use, and Workplace Tracking

Key points

• Senator Ed Markey (Democrat, Massachusetts) introduced four AI accountability bills in July 2025.
• The bills address automated hiring tools, data centre resource use, workplace surveillance, and algorithmic bias in financial and consumer decisions.
• The International Energy Agency projected in its 2024 report that global data centre electricity demand could double by 2026, driven primarily by AI workloads.
• Under the hiring bill, applicants would have the legal right to know when software, not a person, rejected them, and to formally challenge that decision.
• None of the bills attempt to ban AI. They focus on disclosure, limits on data collection, and appeal rights.

Senator Ed Markey of Massachusetts introduced a set of four bills in July 2025 designed to hold artificial intelligence systems, the software behind tools like chatbots, hiring screens, and workplace monitors, accountable for real-world harms happening right now. As first reported by Train2Secure, the package does not try to slow AI development. It tries to shine a light on who pays when things go wrong.

Each bill targets a different part of daily life.

Hiring software that no one questions

Most large employers now run job applications through automated screening tools before any human reads a single resume. An algorithm, a set of rules a computer follows to score or filter candidates, can reject someone without a recruiter ever seeing their name. Markey's hiring bill would require companies to tell applicants when this is happening and give them a formal way to challenge a rejection made by software rather than a person.

A 2023 study cited by the AI Now Institute found that these tools, trained on years of past hiring decisions, routinely penalise candidates from groups that were underrepresented in those historical hires. The bias is not accidental. It is built into the training data.

Water and electricity: costs no one reports

Running AI requires enormous amounts of electricity and water for cooling the servers that do the work. A data centre is essentially a large building full of computers running around the clock. Right now, neither the power draw nor the water consumption appears on any consistent public record.

Markey's proposal would require facilities to publish those figures. The International Energy Agency projected in its 2024 report that global data centre electricity demand could double by 2026. The communities and ratepayers near those facilities absorb that cost today with almost no information about its scale.

Watching workers, second by second

Warehouse workers describe systems that track their movement speed in real time. Call centre staff report software that reads the emotional tone of their voice. Office workers face tools that log keystrokes or measure screen-away time. The third bill would restrict what data employers can collect this way and limit how it can be used to discipline staff.

Bias in loans and financial decisions

The fourth bill targets algorithmic bias, where software trained on historical data reproduces old patterns of discrimination, in areas like loan approvals and fraud screening. No individual human may have made a biased choice, but the outcome is the same.

What does this mean for ordinary people?

For job seekers, it means the right to know a computer screened them out and the ability to ask why. For workers, it means tighter rules on how closely an employer can track their every move. For anyone near a data centre, it means the public would finally see hard numbers on how much water and power those facilities consume.

None of this is law yet. The bills face the full legislative process. But they map onto harms that are already affecting millions of American workers.

"We cannot afford to let Silicon Valley write the rules for a technology that affects every American worker and every American community," Markey said when introducing the package.