What Employers Need to Know About the Use of Artificial Intelligence (AI)

There has been a lot of talk lately about regulating artificial intelligence (AI). How exactly this will be accomplished is currently unknown, for the most part. But certain legal developments are worth noting—including a New York City law taking effect next month—as they are likely the roadmap for things to come.
 
One implementation that has stirred a lot of discussion is the use of human resources AI applications that focus on employee recruitment. These applications can be beneficial as they automate and expedite certain tasks, such as pre-screening interviews, screening resumes, analyzing social media behavior, predicting which candidate will be most successful in a given role, and answering candidate questions about the application process (using chatbots). But these capabilities can also cause harm as these AI applications have already shown a tendency to arrive at decisions that exhibit bias.
 
To counter this, New York City recently became the first municipality to publish regulations that implement an ordinance regulating the use of AI-powered hiring applications (Local Law 144 of 2021). The law applies to companies located in New York City that are hiring candidates for jobs in the city, but it is unclear at this time if it also extends outside the city such as would be the case, for example, if a company outside of NYC were hiring a NYC resident.
 
At a high level, employers who use an AI hiring application must ensure that: (1) the application goes through a bias audit no more than one year before it is used and that the findings are made public; (2) the application is thereafter subjected to an annual independent audit; and (3) certain notices are provided to applicants and employees that such an application is being used.
 
Penalties for violations can quickly add up. First-time violations, and each additional violation occurring on the same day, give rise to penalties of up to $500. Subsequent violations incur a penalty between $500 and $1,500, with failure to provide adequate notice constituting a separate violation.
 
The law goes into effect on July 5.
 
Steps Employers Should Take
 
As noted at the outset, even employers not covered by this law can learn from it and start designing their policies and procedures to reflect what is likely to become law in other jurisdictions. With that, we advise employers to: 

1. Carefully review your agreement with the application service provider. Make sure you understand the scope of protection provided in case the application produces harmful results. If the protection is not enough, point it out and negotiate more favorable terms.
2. Understand the data “pedigree” you are using. Make sure your service provider or data broker uses data that is diverse, representative, and free from discrimination.
3. Be transparent about your use of AI. Inform candidates about how their job application data is being used and allow them to opt out of an AI process.
4. Don’t let AI have the final word. Allow for human judgment, intuition, critical thinking, and soft skills to guide any official hiring decision.

We Can Help
 
Contact Maslon for advice and direction on any questions you have about the use of AI in your company or organization.