Could AI actually be used to manipulate people?

AI could make familiar persuasion tactics cheaper and easier to tailor, such as generating convincing messages or adapting content for different audiences. That possibility is worth taking seriously, but it does not prove that a particular group is running a coordinated psychological operation.

For a specific claim, look for a documented actor, a concrete method, and evidence connecting the method to an effect. A shared name, a funding connection, or a sequence that feels meaningful is not enough on its own. Those can be leads to check, not proof of causation.

I’d separate capability from allegation. A system can generate persuasive text, images, or audio, but that alone says nothing about who is using it or whether it changed anyone’s beliefs. Stronger evidence would include verifiable records of a campaign, clear links between its operators and the content, and independent checks on the claimed reach or impact.

It also helps to compare ordinary explanations. Public recognition, overlapping professional interests, or a name appearing in two unrelated places can look like a pattern without showing a hidden relationship.

For day-to-day purposes, treat highly emotional or oddly personalized messages with the same caution whether they were written by a person or generated by software. Check the claim against more than one reliable source, and don’t pass along dramatic material just because it fits a theory.

That’s a practical precaution, not evidence that a particular story is true. Specific accusations need specific, checkable support.