Artificial intelligence has always been a source of mystery and intrigue, a result of its role in decades of sensationalised sci-fi films—but what is the reality? When AI is used in practical areas like recruitment, is there a real risk? Perhaps, just like any tool, AI is only as good as its data input.
Tomas Chamorro-Premuzic, Chief Innovation Officer at global recruiter ManpowerGroup says, ‘If you train AI on rubbish data, it will give you rubbish insights and recommendations. If you teach AI to copy human preferences, it can replicate and augment human bias and unfairness.’ (Alabi, 2023)
AI Recruitment Tools Increasing in Popularity
During 2023, there has been a surge of interest in AI products like ChatGPT across every industry. In recruitment, this trend has been on the rise for years, with one in three Australian businesses reporting that they were using AI tools for filling positions in 2022. (Ayoub, 2023)
A host of companies are offering AI-augmented tools like chat-based models, one-way video interviews, social media analysers, and automated reference checks. (Barrett & Convery, 2023)
“Managers are increasingly time-poor and looking for ways to shorten the recruitment process. They may be turning to AI to help streamline the administrative processes or to assist in vetting potential candidates,” says Sue Williamson, Associate Professor of Human Resources Management at UNSW. (Ayoub, 2023)
The Risks of Outsourcing Recruitment to AI
Although AI tools seem to be the answer to understaffed and overburdened HR departments, there have been concerning limitations revealed as their popularity grows. Transparency is an issue, as applicants are often unaware that they are participating in an automated process and don’t know how they will be assessed (Barrett & Convery, 2023). And they have reason to be worried, because the factors AI is using to assess applicants have been problematic.
Research has revealed that ‘AI-assessed job applications reinforce biases against women and cultural minorities’ (Ayoub, 2023). The algorithms in these AI tools have been influenced by the prejudices of their human creators, absorbing the patterns of previous managers that rejected candidates for biased reasons. So, the AI then follows this example and excludes those who fit this criteria. (Malik, 2023)
Biased AI Tools Lead to Media Scrutiny and Investigations
Amazon was one of the first companies to have their AI candidate ranking tool exposed. The tool had been trained on resumes from a male-dominated industry, therefore learning that male candidates were preferred and automatically downgrading the women who applied. Resumes that included phrases like ‘women’s chess club captain’ were penalised, whereas verbs commonly found on male applications were elevated. (Barrett & Convery, 2023)
Australia’s Merit Protection Commissioner overturned 11 promotions made by Services Australia between 2021-22 because they were completed without any human review. They have now issued new guidance for businesses in the public sector, covering the myths around AI and revealing the lack of testing and trend of biased decision-making. (Ayoub, 2023)
Regulations and Thoughtful Training Necessary for Unbiased Results
Despite the widespread issues reported, Australia has no laws that govern the use and creation of AI recruitment tools. There is an AI ethics framework, which details principles like transparency, accountability, and privacy, but it’s voluntary. (Barrett & Convery, 2023) There is also no law that says candidates must be informed about AI screening processes. (Ayoub, 2023)
“There are low levels of understanding in the community about AI systems, and because employers are very reliant on these vendors, they deploy [the tools] without any governance systems,” says Natalie Sheard, a lawyer and PhD candidate at La Trobe University. (Barrett & Convery, 2023)
Proven Workforce Management Technology Still the Safe Choice
For AI to be a reliable source of recruitment outsourcing, it needs to be monitored and carefully regulated to ensure that these well-known problematic tendencies can be eliminated.
Once there is a reliable training framework that ensures non-biased decision-making, it could be a useful tool of the future. Until then, the recruitment technology offered by workforce management companies is a proven, human-operated alternative.
References:
Ayoub, S. (2023, October 23). Recruitment by robot: how AI is changing the way Australians get jobs. The Guardian. https://www.theguardian.com/technology/2023/oct/23/ai-recruitment-job-search-artificial-intelligence-employment
Alabi, L. (2023, November 3). Recruiters embrace AI to boost efficiencies and candidate pools. Financial Times. https://www.ft.com/content/87c1bf22-9ad6-4d6a-94e2-270c8633dc8a
Barrett, J & Convery, S. (2023, March 27). Robot recruiters: can bias be banished from AI hiring? The Guardian. https://www.theguardian.com/technology/2023/mar/27/robot-recruiters-can-bias-be-banished-from-ai-recruitment-hiring-artificial-intelligence
Malik, A. (2023, September 25). AI Bias In Recruitment: Ethical Implications And Transparency. Forbes. https://www.forbes.com/sites/forbestechcouncil/2023/09/25/ai-bias-in-recruitment-ethical-implications-and-transparency/?sh=28ad9fc6799f
