AI job interviews border on eugenics, claim experts


Biases in artificial intelligence is an ongoing issue. As AI continues to become entwined in more technologies, its biases start to effect more people. For neurodivergent people, the emerging technology is already causing issues, especially as it becomes a vital element of AI job interviews.

How AI job interviews make neurodivergent lives harder

Reported by the BBC, AI job interviews are becoming more commonplace, and the tools used are far from perfect. Detailed in a BBC Three documentary, Computer Says No, the algorithms used to power modern recruitment tools are bordering on eugenics.

For some companies, these tools are used to automate job interviews. Instead of meeting with someone at a company, a participant will answer pre-recorded answers into their webcam. These answers are then reviewed by an artificial intelligence that governs a participant’s answers and mannerisms.

For neurodivergent people, these tools are quickly become a massive barrier to entry for the workplace. In the documentary, Olly, 24, discusses the issues with AI job interviews and its thin grasp on human mannerisms and how it handles his autism and ADHD.

“It’s quite discriminatory against people like me,” he said. “It goes on your eye contact and gaps in your speech,” he says. “If the computer sees my eyes moving, it denotes that as someone who’s not interested and can’t focus, and can’t engage in conversation for long periods of time. Sometimes I have to slow down and refocus. I feel like I’ve got zero chance of succeeding.”

Read More: Polls suggest Smart Guns with proper safety may finally be accepted by Americans

Current artificial intelligence needs to remove bias

The reason why artificial intelligence penalises neurodivergent people is due to the biases inherent in its dataset. Recruitment AIs are programmed with the idea of a perfect employee, one that doesn't have the mannerisms that some neurodivergent people have.

This issue is not new. As the documentary explains, companies such as HireVue are already removing some of the more controversial features of their algorithms, mostly surrounding facial analysis. However, other companies are not as willing to remove these features.

Nat Hawley, head of community at neurodivergent assistance company Exceptional Individuals, explained that while she isn't against AI, she has noticed its issues.

“I’m not against AI – some neurodiverse people prefer an automated interview,” she explained. “But we need to make sure we’re not using these things before they’re fully formed. I would advise that companies think about how they’re using AI and what the software recognises. The people behind them can train them to be inclusive.”

AI bias is being addressed

The biases inherent in artificial intelligence is being addressed in some parts of the world. Mostly revolving around racial bias — particularly in facial recognition — multiple governments have discussed the issue. In fact, the issue is so prominent that the White House has called for an AI Bill of Rights.

With this in mind, AI could be regulated in a way that could stop it from being a major barrier. Instead, the technology is becoming tired to a eugenics-style barrier, limiting people’s capabilities based on things outside of their control.

For more articles like this, take a look at our AI and News page.