AI usage in recruitment has been on the radar of U.S. federal regulators for a long time, but in recent years, issues surrounding AI usage have gained a lot of traction. Back in 2020, ten U.S. Senators sent a letter to the EEOC (Equal Employment Opportunity Commission) about the use and impacts of hiring technologies and the commission’s ability to conduct much-needed oversight and research on the topic.

In 2022, the EEOC and the U.S. Department of Justice issued guidance to employers explaining how AI can lead to hiring discrimination against those with disabilities, potentially violating the Americans with Disabilities Act.

The guidance document, “Algorithms, Artificial Intelligence, and Disability Discrimination in Hiring”, explains how AI technology tools have been found to be discriminatory, including examples such as job advertisements only shown to targetted groups, game-like online tests that assess skillsets, and video interview software that uses AI to measure speech patterns or facial expressions.

The EEOC and the U.S. Department of Justice warn how AI in recruitment leads to discrimination and can violate the ADA.

Laws being created to regulate AI in recruiting

As AI becomes more prolific within the recruiting space, laws, and regulations are being implemented to safeguard the rights of candidates. Here are a few examples of how lawmakers are reacting to AI recruiting tools:

  • In 2020, Illinois became the first state to regulate the increasing usage of artificial intelligence in recruitment practices, cracking down on the use of AI in video interviews with the introduction of the “Artificial Intelligence Video Interview Act“.
  • Also in 2020, Maryland passed a law that requires notice and consent from candidates prior to using facial recognition technology during a job interview.
  • In 2022, the European Commission proposed regulations to address the use of AI in the EU. The rules will use a risk-based approach, with AI recruitment considered “high risk”.
  • In early 2023, New York City employers will be banned from using AI to screen candidates unless the technology has had a “bias audit” conducted a year before its usage.

Gender and racial bias – a big AI problem beyond recruitment

In 2021, the Berkeley Haas Center for Equity, Gender, and Leadership reported that 44% of AI systems are embedded with gender bias, with about 26% displaying both gender and race bias.

A 2018 study by MIT and Stanford University showed that facial recognition algorithms had a 35% higher detection error for recognizing the gender of women of color, compared to men with lighter skin.

Using AI? Here’s what you need to be doing…

If your company uses AI within its recruitment lifecycle, here are a few things you should consider to ensure you’re in control and in compliance, and you’re providing a transparent hiring experience:

  • Fully understand the algorithms being usedIn the same way there are guidelines surrounding how a candidate is traditionally screened and evaluated, recruiting teams and other stakeholders should be fully aware of the factors being considered by the AI algorithm. Consider all inputs being fed into the screening and evaluating software – is all information job-related? Also, look at the data being created by the system – does it comply with data governance standards?
  • Audit AI tools on a regular basisTools that use AI can adapt to their own findings, meaning algorithms can progress over time. Therefore, teams can’t simply conduct an initial analysis to ensure the results aren’t biased or disadvantageous to a specific group. These tools need to be regularly audited to make sure the AI isn’t unintentionally learning immoral or even illegal algorithms from the data it’s receiving.
  • Understand that outsourced tools don’t eliminate liability risksThere are only a select few companies that have the resources to internally develop AI tools, so most companies use outsourced, third-party recruiting solutions. However, using third-party software doesn’t exempt companies from liable risks, such as allegations of discriminatory hiring practices found within the software. Companies need to ensure recruiters and third-party vendors are compliant with all existing, relevant employment laws.
  • Share how AI is used within your hiring processCommonly, candidates want to know the full ins and outs of the recruitment process – not only to help them succeed but also to build trust. As a matter of transparency, think about what is communicated about the use of AI in the hiring process. Consider informing candidates ahead of time that AI will be used to screen or evaluate their application (in some cases, this could be a legal requirement).

Leading companies taking the high ground

Walmart, Meta, Deloitte, and IBM are some of the leading companies that have joined the newly formed Data & Trust Alliance – an organization that’s helping companies to learn, develop and adopt responsible data and AI practices. Their first initiative is helping companies evaluate recruitment vendors by detecting and monitoring bias in their algorithms.

What are some efficient alternatives to using AI in recruiting?

Using AI within a recruitment process has some beneficial factors, but it can also present a lot of serious challenges and risks. Fortunately, there are ways to create a more efficient hiring experience without the use of AI. We’ve looked at the top reasons teams turn to AI-powered recruiting and provided solutions that offer comparable benefits, without the implications of AI.

Recruiting goal

Solutions available without using AI

Speed up time-to-hire

  • Pre-recorded video interviewing: Evaluate candidates’ pre-recorded interviews at any time
  • Automated scheduling: Real-time availability shared with candidates
  • Automated reference checking: Reference forms are automatically sent to referees
  • Automated communication/reminders: Timely reminders and email/SMS notifications sent to all parties

Improve candidate experience

  • Applicant empowerment: Allow candidates to select how they’re most conformable to conduct their interview – virtual or on-site
  • Automated communication/reminders: Provide consistent communication to keep candidates in the know
  • Mobile first platform: Allow candidates to apply and interview via mobile

Improve the quality of hires

  • Advanced searchability: Mass-boolean search applications and resumes for job-specific keywords and skill sets
  • Pre-recorded video interviewing: Get deep insights into candidates from day one
  • Structured interview methodology: Understand candidates better with a predictive validity of up to 65%
  • Skills testing and proctoring: Assess candidates fairly using proctored skills tests or work samples

Minimize hiring bias

  • Built-in rating guides and rating scales: Evaluate fairly within the interview with HR-approved evaluation tools
  • Structured interviewing: Methodology that assures an equal interview for every candidate
  • Diversified evaluators: Interviews can be recorded and reviewed by a diversified panel
  • Accessibility-friendly system: Candidates can use screen readers and opt-in to utilize other accessibility features

Does AI have a place in recruiting?

AI has incredible potential when it comes to HR and recruiting. The current software has shown it can tackle long-standing, common recruiting challenges by speeding up time-to-hire and eradicating low-value administrative tasks. However, AI software within recruiting is still in its infancy, and because of the technical developments still required, using AI has created its own set of challenges.

It’s likely that regulatory issues and allegations of unfair hiring algorithms will plague AI-powered recruitment software for some time. Companies that use AI in their recruitment processes will likely face severe push-back internally and externally as people become more aware of AI’s presence and power.

AI has a place in recruiting when technological advancements allow the software to exceed the screening and evaluating methods used by human recruiters. It’s unlikely that this level of technology will exist in the next few years, or possibly even in the next decade. Until then, there are many non-AI recruitment solutions available that can offer the same benefits – and more – without the serious drawbacks.

The full blog post was written by Vid Cruiter and published on their website here.