United States

The United States has no single federal law on AI in employment, leading to a patchwork of state and local regulations. The focus is on preventing discrimination, ensuring transparency, and requiring accountability from employers.

State and Local Laws

• New York City, New York: The city’s Local Law 144 is a pioneering regulation. It requires employers using an automated employment decision tool (AEDT) for hiring or promotion to perform an annual independent bias audit and to make a summary of the audit publicly available. Employers must also notify candidates when an AEDT is used, including the job qualifications it will assess.
• Colorado: The Colorado Artificial Intelligence Act (CAIA), effective February 1, 2026, is a landmark law that creates a duty of reasonable care for developers and deployers of “high-risk” AI systems, including those used in hiring. The law’s goal is to prevent algorithmic discrimination, and it requires employers to conduct impact assessments, maintain a risk management policy, and provide disclosures to applicants.
• Illinois: The Artificial Intelligence Video Interview Act requires employers to inform applicants if they will be using AI to analyze their video interviews and to obtain their consent. A separate amendment to the Illinois Human Rights Act, effective January 1, 2026, will broadly prohibit employers from using AI that causes a discriminatory effect on any protected class and will require notification when AI is used for a variety of employment decisions.
• California: California has regulations that clarify that the state’s Fair Employment and Housing Act (FEHA) applies to automated decision-making systems. The regulations make it unlawful for employers to use AI that leads to discrimination against protected characteristics. Employers are also required to maintain comprehensive records of automated decision-making data for at least four years to facilitate oversight.
• Texas: The Texas Artificial Intelligence and Data Act (TRAIGA), effective January 1, 2026, prohibits companies from developing or using AI with the intent to unlawfully discriminate against a protected class. Enforcement is handled by the Texas Attorney General, and individuals do not have a private right of action.

European Union

The EU AI Act is the world’s most comprehensive AI law to date, taking a risk-based approach. Systems used in recruitment and hiring are classified as “high-risk,” which subjects them to strict, pre-market requirements.

Key Provisions

• High-Risk Classification: AI systems used for recruitment, candidate screening, and performance evaluation fall into the high-risk category. This classification triggers a series of obligations for both the developers and the employers (deployers) using the tools.
• Mandatory Requirements: High-risk AI systems must undergo a conformity assessment to verify they comply with the act’s requirements. These requirements include having human oversight, maintaining detailed technical documentation, and using high-quality, bias-free data to train the models.
• Transparency and Accountability: The law requires employers to inform workers and job applicants when they are subject to a high-risk AI system. It also mandates that systems be designed to allow for human oversight and that individuals can appeal an adverse decision.
• Prohibitions: The EU AI Act bans certain AI practices entirely, such as using AI to infer emotions in the workplace.