‘Measure of magnitude' | Using AI for HR? Employees could “opt-out” of AI decision-making under proposed Congress bill

Using AI for HR? Employees could “opt-out” of AI decision-making under proposed Congress bill

The American Privacy Rights Act (APRA), a bill currently in Congress, includes a provision that would require employers to notify applicants and employees when AI is used for workplace decisions and even permit workers to “opt out” of AI use for decisions relating to employment.

If passed by Congress, the APRA would have significant ramifications for how employers can use AI software as it relates to prospective candidates and existing employees, potentially disrupting the plans that many HR teams have for using AI at work.

A survey from Elevate of over 2,900 HR leaders and C-suite executives in the US and Canada finds that over 70% of HR teams are using or planning to use AI in their functions.

The proposal is contained within a bipartisan federal proposal introduced by U.S. Senator Maria Cantwell, Chair of the Senate Commerce, Science and Transportation Committee, and U.S. Representative Cathy McMorris Rodgers, Chair of the House Energy and Commerce Committee.

Although it is still at a very early stage, it is one for employers to pay close attention to given the monumental impact it could potentially have on the use of AI for HR practices.

Opt-out rights & notice: What data privacy bill could mean for AI use

The bill includes two major requirements for employers who plan to use AI for workplace practices. Firstly, the employer would need to give notice to employees and any prospective applicants that AI is being used in a given process, and secondly, it would need to give employees and applicants the right to “opt out” of the use of AI.

The requirements are necessary when employers use AI to “facilitate” a “consequential decision.” This broad scope means this rule could potentially apply to any programs, tools, or platforms used by employers to manage workers that have input from machine learning or AI.

Under the act, “consequential decisions” could potentially extend to decisions on hiring, firing, promotions, performance reviews including performance improvement plans, or any of the myriad applications of AI within the context of HR and employee practices.

But the current level of ambiguity means the bill raises more questions than answers for HR teams. It is not clear yet whether, for example, applicants need to be notified when AI is used to create a job posting or if a company would be prevented from using AI software to assess an entire team or department’s performance if just one worker opts out.

As such, it isn’t yet certain how an “opt-out” would be executed in practice, including whether an employer would need to scrap entire HR platforms and systems or if they would need to provide an alternative option that doesn’t rely on AI.

Looking for more

This means, for example, if a candidate opts out of AI use for their interview and selection process, it is unclear whether the employer would be legally obligated to provide an option for interview, screening, and selection without any AI decision-making, or if they could refuse to interview the candidate.

From work allocation to task automation, AI is already in widespread use with countless potential applications, and if passed, how the APRA is interpreted and enforced could disrupt the plans that many employers have to use such technology. It would also have knock-on effects on the type of products HR tech vendors create and sell.

The APRA says that employers need to give “meaningful information” to applicants and employees on how the AI tool would make the “consequential decision” and the scope of possible outcomes.

Will AI notice and opt-outs really become reality?

HR teams and vendors reliant on AI may be able to breathe (at least a temporary) sigh of relief, however, as the APRA won’t be becoming law any time soon.

It would need to be approved by both committees before reaching the floor and then pass further scrutiny, with critics already in place and similar bills – such as the American Data Privacy and Protection Act (ADPPA) – historically stalling in Congress.

"I'll be carefully reviewing this bill to ensure it doesn't have the same flaws as the failed (ADPPA)," said Senator Ted Cruz. “I trust that a measure of this magnitude will be subject to regular order — i.e., committee hearings and a markup — as lawmakers should expect for any significant tech bill."

Other critics have stated it includes many of the same sticking points as the ADPPA.

Small businesses would also be excluded from the APRA, which presently only applies to companies with an average annual gross revenue for the period of the three preceding calendar years exceeding $40million.

However, HR professionals will need to keep a close eye on the bill, which does have its supporters. Julie Brill, Chief Privacy Officer at Microsoft says the big tech company has been “calling for a federal bill for two decades” and would “welcome the clarity and consistency that a federal legislation would provide,” speaking to the International Association of Privacy Professionals.

Moreover, given the legislation sits deep in a broader data privacy bill, guidance would not necessarily be forthcoming from groups like the U.S. Equal Employment Opportunity Commission or the Department of Labor.

Instead, the Federal Trade Commission, without specific knowledge of employment law, would have two years from the law’s effective date to give guidance, during which employees could file lawsuits against their employer for alleged violations.

Whether or not the APRA does become law, it prompts employers to consider how heavily they wish to rely on AI for decision-making that relates to employees and applicants alike.

You are currently previewing this article.

This is the last preview available to you for the next 30 days.

To access more news, features, columns and opinions every day, create a free myGrapevine account.