Big technology firms are presenting a multifaceted danger to democratic societies, spanning social, political, military and ecological dimensions. Their AI tools are deepening an information crisis that, according to the author, demands an urgent, broad mobilisation. The claim is that these systems function as weapons of mass destruction against democratic processes, eroding trust in institutions and amplifying misinformation.
The piece outlines a chain of harms that begins with the harvesting of personal data and extends to the weaponisation of public opinion, job displacement, curtailment of civil liberties and environmental degradation. Ordinary citizens are portrayed as the primary victims of a predatory economy that aligns its agenda with the interests of the world’s most authoritarian and reactionary states.
According to the author, artificial intelligence is reshaping the relationship between society and truth. For large platforms, the line between factual reporting and falsehood has effectively disappeared, making the reaffirmation of public-interest journalism essential. The stated mission of such journalism is to publish information that serves the public good, to expose propaganda and to sustain a democratic public sphere.
The author argues that AI firms prioritize profit over the public interest, employing methods such as unauthorized scraping of news outlets and the distribution of condensed summaries that divert audiences and revenue away from original publishers. By presenting these extracts as autonomous outputs, the companies allegedly enrich a narrow group of shareholders while undermining the financial viability of independent media.
A clear distinction is drawn between AI-generated text and professional journalism. The latter is described as the product of expertise, ethical standards, source interviews, eyewitness testimony, cross-checking and on-the-ground investigation, often at personal risk. The article cites a recent Mediapart investigation into a far-right political leader that triggered anonymous cyber-harassment and death threats against its editorial team, illustrating the high stakes of factual reporting.
The piece notes that media organisations are legally accountable to courts and readers, whereas AI companies answer primarily to shareholders and often refuse to guarantee the reliability of their conversational agents. It references legal actions by the New York Times, whose publisher cited a $20 million litigation budget, and ongoing proceedings in France through a press-rights collecting society. The author highlights an OpenAI admission that training leading models requires copyrighted material, and a Microsoft vice-president’s comment that quality content improves output, suggesting leverage points for collective pressure on platforms.