It’s like a train wreck; now that I’ve started looking at predictions of where AI could take us in the next thirty years, I can’t look away.

These predictions aren’t just the wild imaginings of science fiction writers. Ray Kurzweil and Geoffrey Hinton have worked in AI for decades. They know where it’s come from, where it’s going, and at what rate.

They empahsize that there’s much they can’t predict — that’s the whole point. We’ll soon enter uncharted territory beyond which no one can know what happens. But they have an idea when we’ll get there, and what kinds of things will be possible.

Medical technology that extends human life indefinitely. Technology that can connect our minds instantaneously to an artificial superintelligence.

Not in a milennium - but within our (previous) natural lifetimes.

AI and the State

Lawyers who work in AI development these days are often engaged in thorny near-term legal issues, such as who owns content generated by AI and whether models will perpetuate discrimination reflected in the data sets they’re trained on.

Those are obviously critical issues that need serious thought immediately.

But when I hear predictions for where we might be in a couple of decades, I can’t help wondering about much more destabilizing legal issues.

This Is the State on AI

For example, if most physical labor becomes mechanized and virtual communication becomes immediate and intimate, will people and goods still move across national borders?

If interactions increasingly (or exclusively!) happen virtually in non-space, can any state claim “jurisdiction”?

And if borders and jurisdiction quickly become outmoded, will we see the Westphalian nation-state simply dissolve in our lifetimes?

For an immigration law scholar, these questions are actually pretty urgent. More on that to come …