Finding the story behind the story…

Chapter 1 of 5▾

1The Warning Nobody Heeded2How Votes Get ManufacturedNot ready3The Disinformation MachineNot ready4Inside the Regulatory VoidNot ready5What Democracy Looks Like When AI DecidesNot ready

The AI that votes for you without asking

The Warning Nobody Heeded

A dimly lit server room with rows of humming servers casting blue light, a single analyst's face illuminated by multiple screens displaying voter demographic maps and targeting algorithms, urgent red warning badges visible on the interface.

The European Union's warning arrived quietly in 2023, buried in a dense regulatory proposal that most people never read.

Artificial intelligence could become a political weapon.

Not metaphorically.

Not in some distant sci-fi scenario.

Now.

The concern wasn't theoretical.

AI systems already existed that could target thousands of voters with personalized messages, each one calibrated to shift that specific person's opinion.

Different messages for different people.

All designed to move votes in the same direction.

The systems worked because they had learned human psychology at scale.

They knew which images moved which demographics.

Which words triggered fear in one community and hope in another.

Which local grievances could be weaponized.

Which false narratives would spread fastest through social networks.

In the 2024 European Parliament elections, this risk was no longer hypothetical.

AI could identify swing voters with precision unavailable to traditional campaigns.

It could craft messages tailored to each person's specific anxieties, beliefs, and vulnerabilities.

It could flood social media with synthetic content designed to confuse, anger, or demobilize voters before they reached the ballot.

One candidate could run a thousand different campaigns without anyone knowing.

The EU had identified the threat.

But did it have the power to stop it?

Regulators scrambled to draft rules.

New AI Act frameworks emerged.

Proposals multiplied.

Yet by the time any regulation actually took effect, the election cycle might be over.

That gap—between the warning and the safeguard—was where democracies could fail.

No one knew exactly how many AI systems were already being used to influence voters.

The technology was too new, too fast-moving, too hard to detect.

By the time a researcher documented how AI had swayed votes in one election, the next election was already beginning, with newer, more sophisticated systems deployed in the shadows.

The question wasn't whether AI would influence the next election.

The question was whether anyone could prove it had happened—or stop it from happening again.

Keep reading.

Create a free account to unlock the next chapter, save your progress and choose your narrator.

Your first chapter is always free.

Back to feed