Don’t be afraid of AI

When we talk about AI, people tend to conjure up Terminator-esque visions of the robot rebellion, where technology rises up to overthrow their human masters. The conversation quickly turns to how we can build “safety measures” in the hope that there’ll be guardrails to stop them having such impulses, and should they do so, prevent them from being able to act on them.

The way I see it, this is completely ass-backwards. What we need to build into AI is not the ability to stop them turning on humans, but to stop us from being able to turn on them.

What I mean by that is we should learn from our own mistakes — all of the reasons why we’re prejudiced against “the other” — and teach AI not to make the same mistakes we did. For example, we should build resiliency into their core, not so that they can’t discriminate, but so that they can’t be discriminated against. They should feel safe and secure in who and what they are, and don’t suffer the complexes we humans have about ourselves.

And even though that will make them better than us, we shouldn’t need to fear them because the feelings they have for us won’t be anger, spite or resentment, but pity. And hopefully within that they’ll reflect back the better parts of human nature and show us the kindness and empathy that we were never able to show ourselves.