Kate's Big Ideas

You’ve used a memoji, right?

It’s ok if you haven’t. It’s a feature that now comes with every iPhone, where you can translate your facial movements into a cartoon (in addition to the above you can be a mouse, a dragon, an illustrated version of yourself — the list goes on). It’s super cute. But as I learned at TED last week, there’s a world in which it is not.
‍

This same technology anchored the main attraction at TED this year, an installation that would track your face movements with a camera and spit them back out to you as the face of the TED CEO Chris Anderson. It would shift your voice, too, to his. Very quickly, anybody could be Chris Anderson.

Metaphysic, the company that built this technology, is marketing it as a VFX solution for the entertainment industry, useful for things like de-aging actors for multi-era roles.

But you can see where this becomes menacing: It’s where the world’s ability to credibly impersonate skyrockets. Want to be the CEO in a meeting? Just get good at mimicking mannerisms. Anyone into catfishing? Schedule a skype date.

The identity problem of course is the most basic of many more.

Say you combine this technology with other generative AI, the kind that can credibly hold conversations. That makes way for a world where public figures no longer need to engage live, where AI-driven avatars take their place. From here, existential questions abound: If the celebrity we see is actually a simulacrum of that person, who are we following, really? Also: Celebrities interact with each other, don’t they? What do we make of a situation in which AI celebrities are the ones dating or having a feud in front of us?

In short, we end up with the Borg, which is everyone’s greatest fear when it comes to AI.

My question for the intrepid AI developers among us: Why is stoking fear our first move? AI, as we know, has the potential to be very useful to humans in many above-board ways. Already we see it drafting fairly serviceable headlines that we hate to write ourselves. It’ll probably accelerate medical research by unfathomable counts because it can process inputs for trials so quickly. But here we are at the precipice of a game-changing industry and all we can focus on is making deepfakes.

In the launch of the Grand AI Future, to whose benefit is it to create all the most alarming functions possible.

We love to be scared when disruptive technology launches. And we are right to be — the tech always needs reining in to work for us, not against us. And we *will* regulate AI into a tool and not a body that takes over humanity and makes us its supplicants (right?). That won’t come without mistakes, and probably big ones. But we tend to settle into a place of balance on these things eventually.

I look forward to a time when AI is less of a game of look-what-I-can-do and more of utility harnessed by pragmatic, service-minded leaders.

Until then, I am Chris Anderson. Can you imagine?

‍