AI Ethics: Taking Stock and the Way Forward

摘要

There is a quote about software programming that goes like this: "If builders built buildings the way programmers wrote programs, then the first woodpecker that came along would destroy civilization." Echoes of this sentiment were felt in the field of AI in 2020 as the aura of infallibility that exists in popular perception about AI faded. As we started wearing masks to prevent the spread of COVID-19, facial recognition systems faltered and provided a hint of the fragility of the underlying intelligence. More consequentially, the dangers of the flaws in facial recognition systems that rely on AI became more apparent as civil unrest and protests happened in different parts of the globe. The concerns about the harms of AI bias were no longer theoretical or confined to academic circles, and major vendors such as Amazon, IBM, and Microsoft suspended or canceled the sale of facial recognition systems in the absence of clear regulations.