Don’t cut humans out of the process.
“AI sounds great, but we need to put humans in the loop,” Matsuoka adds. “And we have to consider that we're building it for humans.” Marcia McNutt, 71, the current president of the National Academy of Sciences, says AI shouldn’t be making decisions for businesses, but instead be “some form of second opinion” that humans can use as aid. Since becoming NAS president in 2015, and re-elected for a second term in 2022, she has directed her organization’s resources towards studies, workshops and research that opens up critical discussion of AI.
Cynthia Breazeal, 55, the dean of digital learning of Massachusetts Institute of Technology, agrees, and says that some of the most important guardrails a technologist can develop are those that encourage responsible use—and, importantly, actively discourage irresponsible use. It’s kind of like writing a moral code–literal moral code–that teaches your model whatto do.