Terrifying images of atrocities start coming in from Country X. Country X has been slaughtering people of Nationality Y. We have all kinds of photographs, video clips, different camera angles, and so on.

It's all so horrifying that Country Y decides to declare war on Country X.

And yet:

The whole thing is fake.

The video footage, the photographs, all of it, is a mere phantom generated by AI.

Will we be able to trust anything coming to us online, or indeed even in video and audio, when AI can mimic not just our voices but also our mannerisms and other idiosyncracies?

That world is coming, and much faster than we think.

And yet there are obvious benefits — quite extraordinary ones, if we're being honest — of AI.

For instance, after returning from my February trip to Spain, I opened up ChatGPT and described to it certain situations that had occurred and what I had said in response to them in Spanish. I know I had been understood, but I wanted to be sure my Spanish had been perfect, and I also wanted to evaluate ChatGPT as a potential tool in foreign-language learning.

In each case, ChatGPT told me my Spanish had been correct, but also offered very subtle improvements. It told me, for instance, that in one case I had used a more formal verb than people normally use in spoken conversation.

I can now get that level of one-on-one instruction and guidance at any time of day or night.