Social media without people: when hate arises from the web itself

In the heart of Amsterdam , a radical experiment has challenged everything we think about social networks. A team of researchers has built a platform populated exclusively by 500 bots , all powered by advanced language models like ChatGPT . No humans. No egos. No personal vendettas. Just algorithms that post, follow, and repost . The goal? To understand whether online hatred and polarization are the product of the human spirit… or the system itself.

The paradox of artificial intelligence

The initial question was as simple as it was provocative: if we remove humans from social media, will the poison disappear? The answer, surprisingly, is no . Even without people, the internet has generated ideological bubbles , amplified extreme content, and focused attention on a few dominant profiles. How is this possible?

Sociologist Petter Törnberg has identified the culprit: not toxic content, but the very structure of the internet feedback loop is triggered that pushes it ever higher. The rest disappears. It's not a question of " bad apples ," but of a basket that rewards the rotten .

Six repair attempts (failed)

The researchers attempted to " fix " the platform with six structural interventions:

• Chronological feed

• Reduction of virality

• Removing Followers and Reposting

• Elimination of bio

• Promoting opposing viewpoints

• Penalization of extreme content

The result? In some cases, worse than before. The chronological feed , for example, has made polarizing posts even more visible. The mechanics reward those who shout, even if the shouters are bots.

The voice that shouts loudest

This experiment forces us to reconsider a widespread belief : that online toxicity is simply a matter of poor manners, angry users, and real-life trolls . But if even emotionless machines replicate the same patterns, then the problem is deeper. The internet itself is designed to reward extremes, to monetize attention, to turn every interaction into a confrontation.

Philosophy of the digital square

The " digital square " is often described as the new agora, the place where ideas, opinions, and worldviews converge. But this experiment tells us that this square isn't neutral. It's a machine that selects, amplifies, and distorts . And with the arrival of artificial intelligence, the process accelerates. It's no longer just users to moderate, but entire factories of polarizing content, mass-generated to capture attention.

The lesson is clear: it's not enough to change the actors; we need to rethink the stage. If we want a network that fosters dialogue rather than conflict, we must address its fundamental rules. It's not just a technological issue, but an ethical, philosophical, and political one. Because the network, like any system, reflects what it rewards. And today, it rewards noise.