Efforts to Upskill AI Work Open Door to Programming Bias Claims

As employers train workers on how to use artificial intelligence, special considerations for women and other historically underrepresented groups in tech fields could trigger liability if it gives them advantages, legal scholars said.

AI’s rapid adoption in the workplace is expected to significantly impact women, who are underrepresented in STEM fields and often work in customer service, human resources, and other industries that have been affected the most by AI, recent studies find.

Amazon, Google, and International Business Machines Corp. are among tech companies that have launched efforts to address the demand for AI talent by equipping workers with the competencies needed to stay ahead of the learning curve. Representatives from these companies didn’t immediately confirm whether any of their US-based programs are exclusively for women.

Despite increased threats from conservatives challenging the legality of certain practices designed to boost workforce diversity, employment law scholars say offering upskilling opportunities to all workers is a legally sound approach. Companies can demonstrate that they’re serious about combating gender disparity in AI without being exclusionary with their practices, which could expose them to costly discrimination claims in court, they said.

The US Supreme Court’s decision last year curtailing the use of race as a factor in college admission practices created a “climate” that views “any diversity initiative as problematic,” said Orly Lobel, a University of San Diego law professor and director of the institution’s Center for Employment & Labor Policy program.

“But practically, I don’t think it’s much of a risk if companies promote gender diversity and invest in upskilling,” Lobel said. “The focus should be on adding diversity in terms of perspective, point of view, understanding the markets, and undervalued talent. It shouldn’t be simply quotas for a certain demographic.”

Jessica Rosenberg, a partner at Akin Gump Strauss Hauer & Feld LLP, advised companies not to base their hiring decisions on workers’ participation in employee upskilling programs. Employers should clarify that these opportunities are only for training purposes, and won’t guarantee or “open the door to a greater ability to be hired within the organization,” she said.

Federal law mandates that employment decisions be based on objective criteria, so considering an employee’s participation in an internal talent development program targeting specific protected groups may lead to allegations of bias, Rosenberg said.

By the Numbers

The demand for upskilling opportunities in AI for women is driven by issues of fairness and representation in data, which risk undermining the innovation potential of AI technologies.

A recent study from the World Economic Forum shows that in 2023 women accounted for about 30% of all AI professionals globally.

While AI talent increased sixfold between 2016 and 2022, female representation grew by only 4%, the report said.

Addressing the root causes of gender disparity in tech is the first step to fostering an inclusive AI landscape that leverages the full spectrum of talent and perspectives to meet business needs, said Judith Spitz, a former Verizon executive.

Otherwise, women will continue to have less influence in shaping the design and deployment of AI technologies, which further exacerbates biases and perpetuates inequalities, said Spitz, founder of Break Through Tech.

The skills-based training program is a joint partnership with Cornell Tech, Massachusetts Institute of Technology, and the University of California, Los Angeles, that aims to create tech career pathways for women and non-binary college students from diverse backgrounds.

“The focus should be on who is in the room” and creating opportunities to tap into historically underserved communities that have the potential to excel but “lacked the privileged access to learn about the technology,” she said.

How to Upskill?

But identifying the workforce gaps doesn’t immediately solve the issue. As employers race to address the widening AI skills gap, they’ve found it challenging to determine the exact skills employees will need to thrive in the future because of the technology’s rapid advancement, industry experts said.

Asha Palmer, senior vice president of compliance solutions at educational tech firm Skillsoft, said this uncertainty necessitates a proactive and continuous learning approach to ensure workers remain competitive. Skillsoft works with companies across various sectors to provide technical training courses to retain and retrain workers.

“It will be cheaper to keep and invest in talent” internally because the skill sets around AI aren’t well established, even as the demand for such workers is at an all-time high, she said.

But the skills required to navigate this continually expanding tech landscape go beyond technical experience, Palmer said. Employers capitalizing on the AI boom are emphasizing critical thinking and other soft skills, competencies crucial for fostering broader problem-solving abilities and adaptability in the current job market, she said.

Public debates have largely focused on AI’s risks, failures, and biases, with little attention on the technology’s ability to improve decision-making across various domains and help understand the root causes of inequality, said University of San Diego’s Lobel, who has written about AI’s impact on the workplace.

This “tunnel vision” focus on AI’s negative aspects and “how bro culture prevails” in the tech sector can create undue fear and dissuade women from considering STEM-related fields, Lobel said.