Updated November 2026: The EU AI Act is now law (Regulation (EU) 2024/1689). The ban on emotion recognition in the workplace has applied since 2 February 2025, and high-risk obligations for HR AI apply from 2 December 2027 following the Digital Omnibus on AI.
In the fast-paced world of talent management, where artificial intelligence and automation are reshaping the HR landscape, ensuring consistent compliance with an expanding and fluctuating set of regulations has become a top priority.
Two groundbreaking pieces of legislation, NYC Local Law 144 and the EU AI Act, are among the most notable regulations to have emerged in this area in recent years. Both laws have profound implications for the HR industry in their respective jurisdictions.
Local Law 144 was officially enforced on 5 July 2023, while the EU AI Act entered into force on 1 August 2024. Its HR-relevant prohibitions apply from 2 February 2025, and its high-risk obligations for employment AI apply from 2 December 2027 (see the implementation timeline).
EU AI Act vs NYC Local Law 144: Scope and applicability
Local Law 144 applies to employers or employment agencies using automated employment decision tools (AEDTs) for candidate evaluation or employee promotion in New York City. It specifically targets computational processes derived from machine learning, statistical modeling, data analytics, or AI that result in simplified outputs used to assist or replace discretionary decision-making in employment.
The AI Act meanwhile applies to providers placing AI systems on the EU market, to deployers (such as employers) located in the EU, and to providers and deployers outside the EU where the system’s output is used in the EU (Article 2). In contrast to Local Law 144, which zeroes in specifically on HR tech, the AI Act covers a broad range of AI systems. But the Act's comprehensive approach means it too applies to systems used in employment-related decision-making processes. Both laws aim to ensure fairness, transparency, and accountability in their respective zones of influence.
Risks, requirements and prohibited practices
All organisations within the scope of Local Law 144 are required to obtain an independent, impartial bias audit for their AEDTs, with the aim of mitigating the risk of pre-existing biases being perpetuated by machine learning-powered employment processes. They are also required to make a Summary of Results publicly available on their website and provide notifications to candidates or employees regarding the use of AEDTs in their evaluation. Unlike the AI Act, the law does not explicitly prohibit certain practices, although the text does specify that the instances of discrimination can be referred to the Commission on Human Rights.
The AI Act outlines four categories of risk, with systems deemed to fall within the highest category, Unacceptable Risk, banned entirely. There is a sliding scale of obligations for systems in the remaining categories. AI systems used for recruitment, selection, promotion, termination, task allocation and monitoring of workers are categorised as high-risk (Annex III point 4). Providers must meet requirements on risk management, data governance, transparency, documentation and human oversight, and undergo a conformity assessment. Employers as deployers must use the systems as instructed, assign human oversight and inform workers and their representatives before use (Article 26). Separately, using AI to infer emotions in the workplace has been prohibited since 2 February 2025, except for medical or safety reasons.
For a breakdown of which employers and employment agencies are covered in New York City, including remote roles and vendor-commissioned audits, see who must comply with the NYC bias audit mandate.
For a sector view with links to every relevant article, see the EU AI Act guide for HR and recruitment. For a California comparison, read the EU AI Act vs California’s employment legislation.
Penalties
In terms of financial penalties, failure to comply with Local Law 144 will land organisations with an initial fine of $500 for their first default and each default on the same day, with fines of up to $1500 for each subsequent default. Every day an AEDT is used without an audit is considered a separate default, while each failure to give a candidate or employee the required notice is also a separate violation. It is important to note that exact fines are determined by the relevant enforcement body or court, and that the law delineates between 'defaults' and 'violations', which carry slightly different penalties. Full details can be found here.
The AI Act meanwhile, whose compliance requirements are more comprehensive than Local Law 144, can lead to penalties of up to €35 million or 7% of global annual turnover for prohibited practices, such as workplace emotion recognition, and up to €15 million or 3% for breaches of high-risk obligations, whichever figure is higher (lower for SMEs).
Reputational and operational hazards
Besides the legal and financial penalties carried by both pieces of legislation, companies using HR in their employment or promotion practices are also liable to suffer reputational and operational damage if they fall foul of the law. If, for example, a company becomes embroiled in a civil lawsuit because their non-compliant use of an AEDT under Local Law 144 leads to an individual or group believing they have been adversely impacted, the organisation's image and brand will likely be tarnished. Human rights actions are also a possibility, with the law not limiting the authority of the New York Commission on Human Rights to enforce its directives. Companies also risk the same reputational harm if they violate the AI Act. On an operational level, non-compliant organisations run the risk of eroding customer trust and disrupting overall business activities. It is crucial for companies to prioritise compliance efforts to mitigate these risks and safeguard their long-term success.
How Holistic AI can help
Holistic AI has been named in CB Insights’ AI 100 list.
- For Local Law 144, our NYC bias audit solution prepares your data, calculates impact ratios and generates the summary of results. The law requires an independent audit, so you may still need a third-party auditor to sign off.
- For the EU AI Act, our EU AI Act readiness assessment classifies HR AI systems by role and risk and maps your obligations.
- The Holistic AI governance platform runs framework assessments for the EU AI Act, NIST AI RMF, ISO 42001 and NYC Local Law 144 in one place.
Hiring in other US states too? See our guide to AI hiring laws by state.



