The death of the internet, disinformation, cognitive warfare: what can be done?

Are we fully appreciating the historical significance of the current transformations taking place on the social web? In this space, which once held the promise of collective intelligence, the proportion of synthetic content (generated by AI or incorporating AI) is reaching considerable levels, however one measures it – newly created pages (71%?), content circulating on social media (25%?), and so on. On another front, and in a longitudinal study, Imperva (Thales) estimated last year that 51% of traffic was generated by bots, whether benign or malicious. This lends credence to the famous ‘dead internet theory’, considered fanciful 5 or 10 years ago, and now openly embraced by Reddit co-founder Alexis Ohanian: ‘so much of the internet is now just dead—this whole dead internet theory, right? Whether it’s botted, whether it’s quasi-AI’. In September, in a remark whose irony (perhaps deliberate) is worth noting, Sam Altman posted on X: “I never took the dead internet theory that seriously but it seems like there really are a lot of LLM-run Twitter accounts now”. Finally, that same month, in a press release, Google contradicted its official line: “the open web is already in rapid decline”.

Yet none of this is at the heart of the debate. Regardless of the social web’s electrocardiogram: we have crossed the boundaries of the social web’s healthy functioning at the same pace as planetary limits, in favour of ‘brainrot’, pure emotion and hypnosis… what some call ‘algorithmic authoritarianism’, driven by states as well as by the lords of this new techno-feudalism. For those who reject the toxic reign of scrolling, the only options left are disengagement or even outright secession due to ‘fatigue’, or migration to the ‘dark forests’—those unindexed spaces preserved from algorithmic logic (messaging loops, community forums, etc.). What is at stake, in the new cognitive infrastructure that the social web has become, is nothing less than the way we live together and perceive the world, the preservation of our social model, and the survival of liberal democracy. This is a defining moment for civilisation.

What can be done? The answers are staggering, whichever angle you look at it from. Technological solutions, such as replacing the GAFAM model with a decentralised Web 5.0. Military and political solutions, such as taking cognitive warfare into account. What role do private players in the fields of communication and public opinion have to play in all this?

In five years’ time, either ‘algorithmic authoritarianism’ will have triumphed, sweeping away like dominoes all the systems of exchange and growth known since 1945, and confronting every player with existential questions. Or we will have freed ourselves from it, and from then on only those brands that have focused on real individuals, on physical spaces of social interaction, on territory and human geography, will retain their strategic advantages.

This translates into a first principle: raising awareness. The intensity of the ongoing transformations of the social web remains poorly understood by many private and public operators, who cling to an outdated image of the web (2.0) as a space for ‘conversation’ where conflict now reigns supreme.

The second principle is ethical in nature: to integrate the fight against information threats as an integral part of CSR – and to draw inspiration from the fight against climate change and pollution, as Fabrice Fries (AFP) suggests.

The third principle is operational and can be implemented immediately: to stop producing these ‘algorithmic ammunition’ formats and feeding the beast. Prioritise meaningful, quality engagement over the logic of virality – which is more fleeting and ephemeral than ever – move away from evanescent KPIs to return to the humanities and social sciences, and shift the focus to people rather than ‘users’. 

Now is the time when we can (still) prepare for this return to reality.