Artificial intelligence and political warfare

Opinion

Hey there, time traveller!
This article was published 03/04/2025 (185 days ago), so information in it may no longer be current.

The end, quite simply, shouldn’t be allowed to justify the means. And the means in this case are downright alarming.

If you’re old enough, you can probably remember doing very, very stupid things as a youngster, without any risk that what you’d done would be forever immortalized on the internet, coming back to haunt you permanently. Photos might exist of you at a cabin party somewhere in the woods that could show extremely bad judgment on your part, but they wouldn’t be found with a simple search on the internet, years later, by a prospective employer.

But the absolute chill being put on internet free speech by the American government with its latest use of artificial intelligence (AI) to strip foreign students of their visas to study in the U.S. — if it goes unchallenged — heralds a whole new level of government oversight and repression.

As Axios reported in early March, the U.S. State Department is using AI in a program called “Catch and Revoke” to review the social media posts of tens of thousands of student visa holders to see if their accounts hold anything that could be considered support for Hamas after its attack on Israel.

If AI finds what it considers to be that sort of support, visa holders receive a letter telling them they’re being stripped of their visa and that they must leave the U.S. immediately or face being forcibly removed from the United States and never allowed to return.

Essentially, leave quickly and quietly without a fight, or risk a lifetime ban from the U.S.

The power to revoke the visas is being credited to the Immigration Nationality Act of 1952, which allows Secretary of State Marco Rubio to strip visas from people whose presence in the United States constitutes a national security risk or hampers U.S. foreign relations, but which has rarely been used for that purpose.

The goal may be defensible in the eyes of the current U.S. administration — halting protests seen to be supporting Hamas — though the foundational right to free speech and the right to protest get trampled in the process.

But an electronic hunt for “fellow travellers” — especially young adults who are only learning the weight and significance of their public opinions — seems like grotesque overreach.

And, if using AI this way is normalized, it looks as though sweeping powers end up being handed to a non-discerning electronic overlord that is simply following programming prompts. It casts a wide, lightning-quick and non-discerning net — and what if the search it’s tasked with next time is for any U.S. federal civil servant who may have spoken ill of a touchy president, or any visitor to American shores who has spoken plainly about their own political beliefs?

What if, when you applied for a government program, AI reviewed your social media to ballpark what it thought your political leanings were before approvals were granted? (Just imagine your fortunes if AI spat out “likely Democrat” under the current U.S. administration.)

Obviously, things posted on the internet are public comment, and if you post, you’re responsible for your words. But AI that can parse reams of data in seconds means anything you say can and will be found, and can be acted on, however small and inconsequential you may feel it has been.

Imagine having to tell your kids not just to be careful about what they post online, but that, quite simply, they should not post any opinion online — because regimes change, and so do a government’s friends and enemies.

And then, in one short generation, free speech will be well and truly dead.