In episode 94 of the AI/Hyperautomation Minute, Aaron Back explains why there must be an intentional focus on ethical AI, especially at the data and modeling layers.
Highlights
00:45 — As part of Microsoft’s layoffs, the company cut its ethical AI team. Although, Microsoft does have its Office of Responsible AI (ORA) that focuses on fairness, inclusiveness, reliability, safety, transparency, privacy, security, and accountability.
01:44 — While ethical AI is necessary for all aspects of AI usage, Aaron focuses on it in terms of the data and modeling layers. At these layers, people are still heavily involved, which has often led to intentional and unintentional bias that impacts the AI output.
03:08 — Human beings can have unintentional bias based on factors such as environment and backgrounds that seep into the AI output and how it’s being used. This can have a real-world impact, like overlooking an applicant for a role within a company because of bias impacting the AI hiring model.
03:47 — Now, there’s even concern around self-repossessing cars. Essentially, if you were to miss a car payment, then an AI algorithm would notice that and prompt the car to drive itself away back to the dealership.
04:09 — With the current hyper-focus on AI, these developments put Microsoft “in a bad light,” Aaron explains, as they seem to be designating ethical AI as “unimportant.”
04:40 — Technology should “aid augmentation and really fuel ideation and creativity,” Aaron describes as one reason why Acceleration Economy will continue the conversation around ethical AI.
