Can AI in recruitment really be objective?

This legislation, known as Local Law 144, will require human resources managers to pre-test hiring tools that incorporate artificial intelligence before using them on job applicants.

It also defines the circumstances under which companies will be required to inform job applicants that they are using such software in their recruitment procedures.

However, critics of the law say its scope is too narrow to regulate most software that incorporates AI.

An earlier version of the legislation stated that companies must verify that their predictive recruiting tools do not replicate any unwarranted bias, according to the online magazine Fast Company.

The move was intended to prevent New York companies from suffering the same setback that Amazon faced in 2018. At the time, the tech giant had to stop using an automated application sorting tool because it systematically discriminated against women.

The final version of Local Law 144 still requires companies to verify that their AI-based recruiting tools are not inducing bias.

But they simply have to make sure they don't discriminate against job applicants based on their ethnicity, gender or nationality.

In other words, this law "doesn’t require vendors to do anything beyond what they were already required to do under federal law," according to Matthew Scherer, senior policy counsel with the nonprofit Center for Democracy and Technology, speaking to Fast Company.

Despite the criticism, Local Law 144 makes New York the first American municipality to regulate the use of artificial intelligence tools in hiring processes.

It is likely that other cities in the United States, but also around the world, will follow suit in the coming months, as companies turn increasingly to this type of software.

According to research by Apec, France's Association for the Employment of Executives, 27% of French companies with more than 50 employees were equipped with these tools in 2017.

A figure that has certainly increased in recent years, as AI has become a bigger part of our lives.

But human resources managers do not always look favorably on the arrival of this technology in companies. Some fear a complete automation of the recruitment process, and possibly the disappearance of their jobs.

However, artificial intelligence is not envisioned as a decision-maker in the selection of successful candidates, but rather as a tool to assist in that decision-making process.

Given the tensions in the labor market, recruiters have even less room for error than before in selecting future employees. They need to be quicker than ever to find new talent, before candidates are courted by competitors or become bored with the recruitment process altogether.

On average, this process can last 32 days in France, whereas jobseekers would like it to last no longer than two weeks, according to a Cleverconnect / Meteojob survey.

As a result, recruiters must innovate to be more efficient and productive when selecting future employees. This is where artificial intelligence tools come into play.

They can be particularly useful at the time of "sourcing", the strategic phase that involves finding candidates that match the requirements of a position, through the posting of a job offer but also the analysis of social networks.

Indeed, AI can quickly scan the resumes and cover letters of potential candidates in search of keywords that indicate specific skills.

This technology is also increasingly used during the interview phase. Companies like HireVue, Cryfe and iCIMS offer software that can assess the motivation of pre-screened talent by analyzing their body language and verbal signals. All without propagating stereotypes.

But opinions differ on that point. Researchers from the University of Cambridge state in a paper, published in 2022 in the journal Philosophy & Technology, that AI-based recruiting tools perpetuate some discriminatory biases and do not improve corporate diversity.

"AI-powered hiring firms’ emphasis on how their tools debias the hiring process deflects from the material inequities and (de)privations that shape applicants’ experiences and profiles.

These inequities cannot be so easily prized apart from the way that AI tools perceive candidates’ behavior, expression of personality, use of keywords, and general self-(re)presentation," note study authors Eleanor Drage and Kerry Mackereth.

While the use of artificial intelligence in recruitment is not yet fully convincing, that doesn't necessarily mean that the technology should not have a place in the process.

However, Eleanor Drage and Kerry Mackereth argue that HR managers need to build up their AI knowledge and expertise to better understand how the recruitment software they will increasingly be using in the future actually works.