Guardian columnist: shame over AI use misses the real target

A Guardian opinion piece published on September 29, 2026 argues that public anger at artificial intelligence is being aimed at the wrong target. Writer Garrison Lovely's column, titled "Stop shaming people for using AI. Start organizing to prevent our obsolescence," makes a direct case: the AI industry is pursuing what he calls "a perverse quest to replace humans," and individuals who use AI tools are not the enemy.

The column opens with a blunt accounting of public sentiment. "I don't need to tell you that people hate AI," Lovely writes, before cataloging the grievances that dominate public discussion of the technology: "slop, cheating, bias, enfeeblement, electricity bills, environmental destruction, exploitation, theft, cybercrime, doom."

That list touches nearly every major fault line in the current AI debate — the flood of low-quality generated content, academic and professional dishonesty, documented model biases, concerns about skill atrophy, the energy and water demands of data centers, labor conditions in the supply chain, copyright disputes over training data, AI-enabled fraud, and catastrophic risk arguments. Lovely does not dismiss these concerns. His argument is about the form the backlash has taken.

The core argument

Lovely's central claim is that resistance to AI has misdirected its energy. "Resistance to the technology has often taken the form of shaming people for personally using it, with precious little organizing against the companies building the machines," he writes.

He finds this pattern self-defeating, and says so in the column's sharpest line: "This is perplexing, given that those companies are attempting to render us obsolete."

The framing matters. Lovely is not making the common argument that AI anxiety is overblown. He accepts the premise that the stakes are existential for workers. His disagreement is tactical: if the technology is being built to outcompete humans, then the strategic response should target the institutions building it, not the individuals adopting it under economic and professional pressure.

The column's subtitle distills the position into a single sentence: "We cannot cede the best tools available to the other side." That line carries the piece's practical logic. If AI systems genuinely represent the most powerful tools available, Lovely argues, refusing to touch them on principle amounts to unilateral disarmament — while the companies developing those systems continue to use them at full force.

The executives' version of the future

Lovely anticipates the counterargument from AI industry leadership and addresses it directly. "The executives may contest this framing, offering some bromides about how AI will augment human work or how we'll simply find new jobs to do," he writes.

He rejects that reassurance. "But their clear goals are to create systems that so clearly surpass our abilities that we will have little hope of competing."

This is the column's second substantive claim, and it is a strong one: Lovely asserts that the augmentation talk from AI companies is rhetoric, while the actual engineering objective is superiority over human capability. The word "bromides" does the work here — he treats the standard industry messaging about human-AI collaboration and labor-market adaptation as comfortable platitudes rather than accurate descriptions of corporate intent.

The argument echoes a long-running tension in AI policy debates. Companies marketing these systems routinely pair claims of rapidly approaching human-level or superhuman performance — used to justify investment and valuation — with assurances that the technology will complement rather than displace workers. Lovely's column refuses to hold both claims simultaneously. If the systems are being built to surpass human abilities, he argues, the augmentation framing does not survive contact with the stated engineering goals.

Why the shaming dynamic persists

The piece's most resonant observation concerns where the social cost of AI backlash actually lands. The grievances Lovely lists — slop, cheating, bias, enfeeblement — are in practice policed socially at the level of individual use. The person who submits AI-generated work, the artist who incorporates generated elements, the professional who leans on a chatbot: these are the visible targets. The companies, capital expenditures, and strategic decisions that shape what the technology can do remain comparatively insulated from that everyday social pressure.

Lovely's proposed redirection follows logically from his diagnosis. If obsolescence is being engineered at the institutional level, then the response requires organizing — collective, structural, and aimed at the entities with the power to decide whether and how these systems replace human labor. Shaming individual users, in his account, consumes the energy that collective action would require, while leaving the industry's actual project untouched.

The column does not specify what that organizing should look like — the excerpted portion available does not outline specific tactics, policies, or institutional demands. What it establishes is the target and the rationale. The piece functions as a provocation to reorient, arguing that the choice between "use AI" and "oppose AI" is a false one when the real question is who controls the technology and toward what end.

The stakes

The column lands at a moment when the terms of AI criticism are actively contested. Environmental advocates focus on data center power demands. Creators pursue copyright litigation. Educators rewrite assessment norms. Labor organizers track displacement in specific sectors. Lovely's argument cuts across all of these: he treats them as symptoms of a single underlying project — the deliberate construction of machine capabilities that exceed human ones — and argues that a backlash fragmented into personal moral judgment cannot answer an institutional program.

The piece's own words frame the stakes plainly. The industry, Lovely writes, is engaged in "a perverse quest to replace humans." His answer is not abstinence from the tools but a refusal to accept the terms on which they are being developed. "We cannot cede the best tools available to the other side" is both the column's closing argument and its practical wager: that organized pressure, not individual virtue, is what will determine whether human labor remains competitive as these systems advance.