Protecting workers' rights while using AI

The US experience

In the US, there have been a number of lawsuits in the last few years related to discrimination against employees due to the use of AI solutions. For example, in 2022, iTutorGroup was found to have violated the Age Discrimination Act in the US: its AI-based recruitment software rejected more than 200 older applicants. As a result, the company paid $365,000 to candidates who were rejected because of their age.

That is why the trend of regulating the use of AI and related technologies in automated employment decision tools is rapidly developing in the United States. As of this February, such bills were being debated in the U.S. Senate in New Jersey and New York, in Massachusetts,¹ and in Illinois.

Regulation is centered on the use of an Automated employment decision tool (AED). The AED tool is based on AI and related technologies. The tool automatically filters potential candidates for employment or regulates labor relations with employees.

The main regulatory goals are (1) to prevent workplace discrimination in the AI use and related technologies; (2) to ensure human control over AED tools and its solutions; and (3) to ensure transparency and understanding of the operation of AED tools.

The following areas that are emphasized in the regulatory framework can be highlighted:

1) Mandatory notification to the candidate/employee that the AED tool is being used.

2) Prohibiting or restricting the technology use that can read emotions or analyze human behavior.

3) Imposing human control over the generated decisions. For example, in Massachusetts, an employer cannot rely entirely on decisions generated by the AED tool, especially when determining employee wages and deciding whether to hire, promote, terminate, or discipline employees.

4) Conducting an annual audit of AED tools. Such audit is aimed at preventing discrimination of employees, including age, race, etc., and at identifying errors, deviations, violations of employees' rights.

5) Introducing special regulation for employers who ask candidates for employment to record video interviews and analyze such videos with AI or use facial recognition services.

Russia’s experience

Currently Russia lacks regulation of AI in hiring, although similar technologies are common in the country (e.g., using a robotic recruiter (“Vera”).

In Russia, it may be recommended to supplement the Labor Code (197-FZ) Article 22.4, securing the right of an employee or candidate for employment to know that the employer uses automated decision-making tools based on AI or other similar technologies, to establish a prohibit on the use of such technologies for the purpose of discrimination against an employee or for purposes unrelated to labor. An audit of automated technologies is also important.

From the monitoring issue No. 2, February 2024. Download the full issue (PDF) · issue page at the Gaidar Institute