As experts warn of the discriminatory risks of relying too heavily on AI in recruitment, job seekers are venting their frustrations as their applications are rapidly being rejected.
Rehan Morell, a Victoria-based human resources expert, said in a LinkedIn post last week that she suspected a company was using AI in its recruitment process after her job application was rejected within two hours.
“I’m really shocked and confused,” she said. “I applied for a position at 1:13pm today, and now it’s 3:17pm and I just got rejected from that position. Will the recruiter be able to consider my application along with all the other applications and reject me within 2 hours?”
Morell said her experience was “perfectly” aligned with the job description and it was clear the business was using AI.
“In fact, I have more experience than the role requires,” she said in the post. “I think it took me longer to write the application than it did to get it rejected. This is a new record for my rejection. It used to be 5 hours.”
Companies are increasingly relying on AI throughout the hiring process, from scanning and screening resumes to conducting unsolicited chats and video interviews.
Sapia, a long-standing AI interviewing platform, has conducted 9 million AI interviews for clients including Qantas, Woolworths, Bunnings and other major employers. The company’s interviewing platform improves the experience for hiring teams and candidates, and claims a 95% job seeker satisfaction rate.
Interviewees will have a text chat with an AI agent. The AI agent provides details about the company and job title, explains how chat data will be used, warns against copying or generating text, and asks five questions such as “Describe a time you overcame a challenge” or “Where do you see yourself in five years?” – Interviewees are encouraged to respond using 50-100 words with no time limit.
Barb Hyman, founder of Sapir, said the results from the platform are more fair because the platform does not collect demographic information such as age, gender or appearance, only from interview responses.
“There is a huge amount of bias in hiring, especially in mass hiring,” she says. “Sapir is built on fairness. Anyone can be interviewed and share who they are, in their own words, on their own time. There’s no personal data. It’s a truly blind and fair way to evaluate people.”
Australian HR Association chief executive Sarah McCam-Bartlett said while some candidates were frustrated by increased use and screening of AI by employers, AI was also being used deliberately by companies seeking to reduce bias.
“Interestingly, some employers are now using AI to strip personal information from resumes and applications, so that decision-makers don’t exhibit unconscious bias based on gender, where a candidate lives, age or ethnicity,” she says.
However, while Sapia has “very mature” customers with a sophisticated understanding of how to use AI to make responsible decisions, Hyman acknowledged that many customers are still new to the technology.
“I think 90 percent of the market is just buying without necessarily requiring that level of scrutiny,” she says.
Australian Services Union national secretary Emmeline Gasquet said there was good reason for workers to be deeply concerned about the role of AI in the employment process.
“We know that algorithms are full of bias and prejudice, which is why we need human judgment in hiring decisions, not just machines.”
Connie Jen, an associate professor of human resources management at the University of Adelaide who studies the use of AI in recruitment, says there is still a clear need for human oversight and legal guardrails.
“We have found that organizational guidelines and legal requirements such as non-discrimination HR policies are more effective. [than AI] In addition to contributing to improving diversity and inclusion, it was also important to have a human resources manager with a focus on diversity. “We found that AI doesn’t make a huge difference,” she says.
A study by University of Melbourne lawyer Natalie Sheard found that when employers use AI recruitment systems to screen and shortlist candidates, they risk engaging in “algorithm-facilitated discrimination”.
The limited data used to train these systems can not only fail to reflect the diversity of the population and entrench traditional forms of discrimination, but also create new forms of discrimination and pave the way for intentional discrimination, she says. Sheard said discrimination facilitated by algorithms is particularly problematic because the predictions and results produced by these systems are often contested, and the processes used tend to be opaque.
The federal government has committed $30 million to establish an AI Safety Institute to monitor, test and share information about new AI uses, risks and harms, and in December introduced a National AI Plan that sets out voluntary guardrails for the deployment of generative AI.
But Sheard says the government needs to review and reform its discrimination laws to adequately protect job seekers.
“If we do not want disadvantaged people to be subject to algorithm-facilitated discrimination, we need to take urgent action,” she said.
Get workplace news, advice and perspectives to help you get ahead at work. Sign up for our weekly Monday newsletter.
Millie Mulloy is an economics writer for the Sydney Morning Herald and The Age. She was previously an economics correspondent based at the Press Gallery in Canberra and a banking writer based in Sydney.connect via Twitter or Email.
Tim Biggs is a writer covering consumer technology, gadgets, and video games.connect via Twitter or Email.
