We may create fantastic machines, but will they absorb our own flaws?
Our future is bright, but it seems there are some problems we can’t shake off. Sadly, sexism might be one of them.
The newly appointed UK Artificial Intelligence leader at PwC has warned of the risks of tech being designed by overwhelmingly male teams.
There has been growing concern – especially after AI systems used by US police were shown to discriminate suspects based on race – that gender bias will soon become an issue too.
“One of the obvious issues in the AI environment is gender diversity that’s characterised by more males than females, particularly on the technical side, and which can lead to potentially hidden or unconscious biases being coded in,” Euan Cameron told The Memo.
Even if a diverse team designs an AI, says Cameron, there’s no guarantee it won’t absorb biases when it is being trained or learning from a human.
For instance, even among the most diverse groups of people, studies have shown words like “female” and “woman” are more often associated with jobs in the arts and humanities, while “male” and “man” are closely associated with maths and engineering.
It’s also been proven that AI is picking up these subtle gendered associations.
Cameron says he’ll be working with companies to highlight these biases and recommend ways to protect against them.
“You need a process of governance around an AI system so at least you can take every step to make sure you have guarded against bias becoming built in.”
It’s a big job and one that’s only set to get more pressing over time as artificial intelligence spreads through our society.
For his part, Cameron predicts that there’ll be no corner of our world that won’t be touched by the new technology.
“It’s a quiet revolution, which is not immediately obvious to the users or consumers, but will nevertheless have a fundamental impact on the business models across every sector.”