Innovating Through AI For More Equitable Talent Outcomes

With the game changing technological innovation of generative artificial Intelligence (GenAI), companies will require more than incremental innovations—many business models will be entirely transformed.

The accelerated adoption of generative AI tools like ChatGPT has brought new focus to AI’s potential in talent processes. However, there are differing opinions regarding AI’s long term impacts on diversity, equity, and inclusion (DEI). While some leaders are concerned that AI has a propensity to perpetuate bias at scale, others believe AI can effectively accelerate de-biasing efforts. At a time when DEI-specific recruiting and development programs are being challenged in the US courts, leaders can leverage AI to accelerate equitable talent outcomes.

Although executive recruiting is very high touch and somewhat immune, workforce talent processes can be transformed through AI. For example, in talent acquisition, applications of AI include automated sourcing of online profiles and cataloging of skills and experiences; automated resume screening to identify profiles that best fit the criteria in job specifications; and candidate assessment via AI-supported analysis of interview transcripts and sentiments. In talent development, AI can be used to analyze workforce performance data and generate predictive success profiles. These applications can result in both increased efficiency and effectiveness in talent processes.

That said, the potential for bias to creep into both data sets and decision-making algorithms is a major risk when using AI in talent systems. For example, some tools utilize HRIS (human resources information systems) company level data to identify who has been successful in jobs based on performance data analysis and individual profile data, found on sites like LinkedIn. If the HRIS workforce performance data includes a gap in promotion rates between certain demographic groups (e.g., men vs. women), the existing bias represented in the company data may be perpetuated when using analytics to predict successful profiles.

However, bias can also be disrupted. These mechanisms can be built into data sets, as well as the algorithms analyzing them. In the example of combining external profile data with company performance data to generate predictive success profiles, organizations can preemptively correct the promotion differentials that may exist between demographic groups within the algorithm. When applying AI to screen resumes against job specifications, bias disruption techniques might include scoping the job description to include only “must have” vs. “nice to have” job requirements, which opens the aperture and optimizes potential applicant diversity.

Traditionally, DEI programs have been designed for under-represented demographic groups to increase their representation at every organizational level. But in the aftermath of the US Supreme Court decision on race conscious admissions, DEI programs like these are being challenged. Additionally, many organizations are finding that siloed DEI departments can only have limited impact.

Modern approaches embed equity practices into the full talent management life cycle. For example, in talent development, a desired outcome could be to ensure that all demographic groups are adequately represented in overall leadership development programs. Organizational leaders can embed equity steps into the sourcing process by leveraging AI to: identify a wide range of participants, efficiently scan the relevant employee population, systematically assess applicants by matching assessment data with future potential roles, and tabulate individual calibrations to provide a selection recommendation that is not influenced by group think. As leaders seek more equitable talent outcomes, AI is a powerful tool for human capital innovation.