Ethics Boards, Codes of Conduct Help Guide AI Use

By Dave Zielinski March 5, 2019
Ethics Boards, Codes of Conduct Help Guide AI Use

​Emerging alongside reports of the many ways artificial intelligence (AI) is benefiting HR is a more troubling narrative that algorithms may perpetuate bias in hiring or other talent decisions.

While well-designed AI can help eliminate unconscious human bias in processes like candidate screening, there also are risks that the technology can cause adverse impact as well. The bombshell revelation last year that Amazon created and later shut down a recruiting algorithm that was biased against women applying for software development jobs continues to reverberate in HR. Although it represented just a single case of an internally developed AI tool, the flawed Amazon algorithm captured the attention of the industry and led to a new wariness and some policy changes.

Some organizations have taken these ethical concerns seriously enough that they've created advisory boards and new codes of conduct to monitor and guide their use of AI. HireVue, a provider of AI-driven video interviewing and assessment tools in Salt Lake City, has created a board to guide the ethical development and use of AI in its products. The board includes experts in industrial and organizational psychology, algorithmic bias, data privacy, and diversity and inclusion.

Loren Larsen, chief technology officer at HireVue, said the company had been watching the increasing pace at which its AI applications were impacting job seekers and believed it was time to formalize its code of conduct for using the technology. Among other uses, the company's AI is programmed to notice factors in video interviews like facial expressions, what candidates say and how they say it.

"We wanted to send a signal to our customers and to the market that we care about using AI technology ethically, because there are ways it can be used irresponsibly," Larsen said. "We decided to create a board of experts to help ensure we're not overlooking any ethics issues, as well as remaining state-of-the-art with our technology."

Weston, Fla.-based Ultimate Software also has created a code of ethics for its use of AI, said Armen Berjikly, the company's senior director of growth strategy. "The key is to not simply let technology 'lead' and deal with the ramifications later, as so often happens when we have new capabilities," he said.

"In HR, we are dealing with people's livelihoods. We need to be proactive as an industry about where we want AI to go and conversely where it should not."

[SHRM members-only platform: SHRM Connect]

A Starting Point

Experts say creating advisory boards and codes of conduct can be a positive first move to ease the minds of consumers of AI tools and assure them steps are being taken to protect against bias or abuses of the technology.

"I think it pays to be very transparent about how you're approaching your use of AI and machine learning these days," said Ben Eubanks, author of Artificial Intelligence for HR (Kogan Page, 2018) and principal analyst at Lighthouse Research and Advisory, an HR research and consulting firm in Huntsville, Ala.

One of the biggest concerns HR and recruiting leaders have about AI applications is making sure they evaluate job candidates and employees without bias. "It boils down to [asking] are people being treated fairly and are they being measured accurately," Larsen said. "We want to ensure we're not creating any bias in terms of gender, race, age and other factors in the way we use our technology to conduct video interviews and assessments."

Jeanne Meister, founding partner of Future Workplace, an HR advisory and research firm in New York City, believes organizations shouldn't rely only on vendors' AI codes of conduct or advisory boards. "I think it's encouraging that some vendors are doing this, but ultimately companies need to create their own code of conduct for using artificial intelligence in HR and communicate it to all employees," she said. "Some recent studies have pointed to how few companies actually communicate to the workforce about how their algorithms are created, what the assumptions are behind them, and where the data generated from them is stored and how it is protected."

How to Assess Your AI

Eubanks said HR leaders are responsible for ensuring that algorithms they employ from vendors don't create any adverse impact.

"You can't be satisfied with the response of 'it's proprietary' from a vendor when you ask about how their algorithms work because that's not good enough," he said. "At the end of the day, if you start making biased decisions about who to hire based on an algorithm from a software vendor, the worst thing that can happen to that vendor is you'll fire them. But the worst thing that can happen to you as an HR leader is you may lose your job and affect your career because you'll be held accountable as the person cheating candidates out of jobs for using a biased AI product."

Eubanks said HR leaders should conduct thorough due diligence on how algorithms they employ are validated and tested, as well as on their track record of predictive accuracy.

"If someone tells you they have a proven way to predict future performance, ask them about the specific performance factors they're testing for and their results," he said. "It never ceases to amaze me, for example, how many startup companies I talk to in the talent acquisition space have never actually done any recruiting. They are very smart, technical people, but they don't always have experience in HR."

Eubanks said companies like IBM have created new processes to protect against algorithmic bias. "I recently spoke to a chief engineer at IBM who helps develop algorithms for Watson," he said. "She said one of the things they do is make sure at every single step of the algorithm development process a diverse team is involved in the creation. They believe if algorithms are designed by a homogenous group, all the potential biases and outcomes of the AI won't be considered and mitigated."

Dave Zielinski is a freelance business writer and editor in Minneapolis.


Job Finder

Find an HR Job Near You
Search Jobs


HR Daily Newsletter

News, trends and analysis, as well as breaking news alerts, to help HR professionals do their jobs better each business day.