Regulating and controlling the powerful and wide-ranging technology of artificial intelligence will require a complex response that includes ways to manage risks such as misinformation and deepfakes, experts say.
AI is powered by complex algorithms, which are essentially mathematical equations. They sometimes have billions of different parameters and don’t produce the same outcome twice.
But the algorithms, which help determine everything from the ads we are shown to the jobs we are considered for, can sometimes produce concerning results. AI’s potential to reinforce discrimination is one of its most worrying effects.
Critics point to an oft-quoted example of a failed Amazon algorithm that was trained to analyse job applications based on a data set of 10 years of successful job candidates. Most hires had been men, so the system only put forward male candidates, consequently reinforcing gender bias in the business.
Read the full article in The Australian Financial Review.