published by xavier.grehant on 2026-05-12

The Epistemic Boomerang: Will AI Destroy the Information Media It Accelerates?

I. The Supply-Side Catastrophe

The information ecosystem has always been vulnerable to manipulation, but generative AI has removed the last remaining friction. Before the internet, mounting a disinformation operation required capital or manpower. Today, as researchers at the Bulletin of the Atomic Scientists observe, AI folds that overhead into relatively cheap keystrokes, producing what some warn is a "chain reaction of harm" with the potential to deepen public health crises, hinder disaster responses, and undermine democracy — with NewsGuard counting more than 1,200 AI-driven unreliable news sites by May 2025. The gates have not merely been opened; they have been dissolved.

This is not simply a quantitative expansion of old problems. The advancements in synthetic media have reshaped the very perception of audiovisual content, challenging its traditional role as a trusted source of evidence. A deepfake is not just a lie — it is a weapon against the epistemological contract that underpins all media: the assumption that recorded reality is at least a starting point for truth. When that contract breaks, every document becomes suspect. The World Economic Forum ranked AI-generated mis- and disinformation as the most significant short-term global risk in 2024, surpassing even the threat of extreme weather events. The concern is existential in scope.

Journalists themselves feel the tremors acutely. A structured survey of 504 journalists in the Basque Country found that 89.88% believe AI will considerably or significantly increase the risks of disinformation, with this perception more pronounced among those with the greatest professional experience. The veterans — those who built their careers on the credibility of the medium — are the most alarmed.

II. A Historical Parallel: The Pamphlet Wars

This is not the first time a new reproduction technology threatened to dissolve the epistemological order. When Gutenberg's press democratized the written word in the 15th century, it unleashed a century of pamphlet wars, apocalyptic prophecy, and theological disinformation on a scale Europe had never seen. The Reformation was as much an information crisis as a spiritual one. Likewise, the penny press of the 19th century — and later Hearst's yellow journalism — weaponized the mass printing press for sensationalism and war-mongering, culminating in the Spanish-American War of 1898, partially manufactured by headlines. Each time, the medium that accelerated communication also destabilized it. And each time, society adapted — through literacy, regulation, professional codes, and institutional gatekeepers.

The question today is whether the adaptation cycle is fast enough, and whether the scale of AI-generated content has crossed a qualitative threshold that renders historical analogies insufficient.

III. The Paradox of Scarcity: Credibility as a Premium Good

Here the picture grows more nuanced. A counterintuitive dynamic is emerging from the data. While exposure to AI-generated misinformation makes people more worried about the quality of information available online, it can also increase the value they attach to outlets with reputations for credibility, as the need for help distinguishing between real and synthetic content becomes more pressing.

This was confirmed empirically in early 2025. Researchers examining 17,000 readers of the Süddeutsche Zeitung found that AI-driven misinformation lowered general trust but also raised engagement with trustworthy news sources. The logic follows classical economics: when a good becomes scarce — in this case, verified truth — its perceived value rises. AI may inadvertently be creating a premium market for rigorous journalism, not destroying it.

That said, this dynamic is highly unequal. AI-enabled misinformation largely succeeds with those who already agreed with the broad intent of the false message, while leaving skeptical others unconvinced. The information landscape may therefore bifurcate: a small, educated stratum retreating to high-trust institutions, while a broader population sinks deeper into algorithmically curated unreality. This is not a stable equilibrium — it is a social fracture.

IV. The Analog Boomerang

Beyond the battle over institutional credibility, something more visceral is happening. A generation raised entirely on digital media is staging a cultural retreat. A 2024 Pew Research study found that 62% of adults feel "overwhelmed" by digital life, craving moments of disconnection — turning to vinyl records, film cameras, and physical journals as artifacts of intention and permanence.

The numbers are not marginal. The Vinyl Alliance's 2025 report showed that 50% of Gen Z vinyl enthusiasts see the medium as a form of "digital detox," while 61% aim to improve their well-being by shifting away from digital consumption; the percentage of 12-to-15-year-olds taking deliberate breaks from smartphones has risen by 18% since 2022. Print is holding. Physical books still account for the majority of book sales, and bookstores are filling up again — with the revival described not as nostalgia but as a shift in how people value presence and long-form attention.

In marketing, the signal is equally clear. A 2025 Harris Poll found that 71% of consumers believe print catalogs or magazines feel more authentic than digital ads, and the Global Wellness Institute identified "Analog Wellness" — the mainstreaming of pre-digital hobbies and tangible rituals — as a top trend in its 2025 report.

This is not a rejection of technology per se, but a renegotiation of its terms. The medium that claimed to bring the world closer has, for many, made the world feel less real. When reality itself is rendered uncertain by synthetic media, direct physical experience becomes the last epistemological refuge.

V. Could AI Kill the Media It Accelerates?

Not in a single catastrophic collapse, but through a slower erosion that reshapes the media landscape beyond recognition. Three dynamics are converging:

The pollution of the commons. When AI-generated content floods digital channels at scale, the cost of truth-verification rises for everyone. Readers either develop costly discernment habits or abandon the effort entirely. Neither outcome benefits open media.

The bifurcation of trust. High-trust institutions may survive and even thrive as scarcity premiums rise. But the middle — the regional newspaper, the independent digital outlet — faces existential pressure, already weakened by revenue collapse and now further undermined by ambient distrust.

The retreat to the tangible. As digital space becomes epistemologically hostile, people will increasingly anchor their reality in direct experience, physical objects, live events, and face-to-face community. This is not regression — it is a structural response to informational overload and manufactured unreality.

History suggests that communication technologies do not die — they find their ecological niche. The telegraph did not kill the letter; radio did not kill print; television did not kill radio. But each forced a redefinition of purpose. AI may do the same to digital media: strip it of its claim to be a reliable mirror of reality, and force it to reinvent itself as something else — a space for conversation, entertainment, and opinion, no longer confused with testimony or fact.

Conclusion: The Return of the Real

There is a profound irony at the heart of the AI information crisis. The technology that promised to give us more knowledge, faster, may end up producing a civilization more skeptical of mediated knowledge than any since the Enlightenment. Not because people become wiser, but because the noise overwhelms the signal until direct experience — touching, witnessing, being present — becomes the only form of evidence people trust.

Jean Baudrillard, writing in Simulacra and Simulation (1981), warned that Western society was already replacing reality with representations of reality, until the map preceded the territory. AI is his argument made operational. The question is whether the human instinct for the real — for the smell of paper, the weight of a handshake, the unrepeatable singularity of a live concert — will act as an antibody. The evidence of 2024–2025 suggests it might. The boomerang, thrown hard into the digital ether, may be arcing back toward ground.

Sources

Bulletin of the Atomic Scientists (2025)
CEPR / Campante, Durante, Hagemeister & Sen — GenAI Misinformation, Trust, and News Consumption: Evidence from a Field Experiment (2025)
World Economic Forum — Global Risk Report (2024)
CMU Heinz College / Süddeutsche Zeitung field experiment (2025)
Springer Nature AI & Society — scoping review on generative AI and misinformation (2025)
Pew Research Center (2024)
Vinyl Alliance Report (2025)
Global Wellness Institute — Future of Wellness (2025)
Harris Poll (2025)
MIDiA Research — Analogue Revival: A Cultural Pendulum Swing (2024–2025)
Jean Baudrillard — Simulacra and Simulation (1981)