Do you trust A.I.? No? May I ask why not?
Self-driving car crashes, you say? Automation job-killing apocalypse? A complete takeover and destruction of humanity by rogue super A.I.?
Well consider this missive, from Activa co-founder and CEO Rana el Kaliouby:
“A.I., doesn’t trust us either.”
She made this astounding statement in her keynote address at the second Emotion Ai Summit, held in Boston, Massachusetts on September 6. Trust in A.I., was the theme of this year’s meeting, and with good reason. The meeting covered the ethical and trust issues in A.I., in areas as diverse as autonomous vehicles, product marketing and education.
Since last year’s inaugural summit, which was held by Activa at the iconic MIT Media Lab, the news has been full of not-so-encouraging stories about a possible dark future of A.I. More than one economic pundit has predicted a massive kill-off of jobs by smart automated systems. And Elon Musk and, until his recent demise, Stephen Hawking, have been all over the media with warnings of an A.I. doomsday.
So, what’s with Kaliouby’s position? As the CEO of perhaps the foremost producer of emotion-savvy A.I. software, she obviously has motive to persuade us to trust AI. But why wouldn’t it trust us?
Perhaps the statement was hyperbole. She explained it as the need for A.I. to trust that it is getting good input from us, so it can make the right decisions. But until we have sentient, general A.I., it might better be interpreted this way. To trust A.I., we first must trust ourselves to provide the right programming and input for A.I. As one presenter put it, the goal should not be to create good A.I., but A.I that does good.
In her closing address, el Kaliouby put forth what she called a three-part contract with A.I. Trust—mutually—is the first part. The second part is pervasiveness. She feels it needs to ultimately encompass virtually all our experience. And third, it needs to be ethical; this assumes we can define what that is.
But perhaps the most telling comment came from one member who appeared on a panel of venture capitalists who discussed investing in A.I.
When asked what is it that excites you the most and that scares you the most about A.I., Habib Haddad, of venture fund E14, said his greatest worry is, “what will kill us first, artificial intelligence or natural stupidity?’