Pretty much, but the scary thing is that it is learning what people want to hear and in the process often amplifies human bias. It picks up on things like racism and uses that in its decisions. Then those decisions go into the records that AI systems are being trained on, entrenching the biases more and more over time. Because the processes are pretty much invisible it is hard to do anything to stop that from happening. Then we have this weird idea that computers are somehow impartial and should be trusted over human decision makers, and we are using AI to make decisions on things like who to shortlist for employment, who to flag for closer scrutiny in immigration, what level of aged care someone receives - all major life changing things!
Pretty much, but the scary thing is that it is learning what people want to hear and in the process often amplifies human bias. It picks up on things like racism and uses that in its decisions. Then those decisions go into the records that AI systems are being trained on, entrenching the biases more and more over time. Because the processes are pretty much invisible it is hard to do anything to stop that from happening. Then we have this weird idea that computers are somehow impartial and should be trusted over human decision makers, and we are using AI to make decisions on things like who to shortlist for employment, who to flag for closer scrutiny in immigration, what level of aged care someone receives - all major life changing things!