Prominent science-fiction author and activist Cory Doctorow has issued a stark warning: the current artificial intelligence boom is a speculative bubble destined to collapse, leaving a damaging legacy he compares to 'asbestos in the walls of our technological society'.
The 'Reverse Centaur' Workforce
Doctorow introduces a key concept for understanding AI's real-world impact: the 'reverse centaur'. Unlike a 'centaur', where a human is augmented by a machine, a reverse centaur is a human forced to serve as a 'squishy meat appendage' for an AI system. He cites the example of Amazon delivery drivers, monitored by AI cameras that penalise them for looking in the wrong direction or singing, effectively making the driver a peripheral for the van.
'Obviously, it’s nice to be a centaur, and it’s horrible to be a reverse centaur,' Doctorow states. His thesis is that many AI tools are created not to empower workers, but to create these exploitative reverse-centaur roles, where humans bear the accountability for AI's mistakes without reaping the benefits.
Monopolies and the Growth-Stock Crisis
Doctorow roots the AI investment mania in the structural crisis of tech monopolies. Giants like Google, Meta, and Apple dominate their sectors, but face a Wall Street paradox. As growth stocks, their valuations rely on perpetual expansion. Once growth stalls—as when a company achieves a 90% market share—they risk a catastrophic re-rating from 'growth' to 'mature' stock status.
He points to Meta's experience in early 2022, when news of slightly slower user growth triggered a $240 billion single-day sell-off. To avoid this fate, dominant firms must constantly hype new, transformative technologies—whether the 'metaverse', cryptocurrency, or AI—to maintain the illusion of limitless growth. 'The primary goal is to keep the market convinced that your company will continue to grow,' Doctorow argues.
Why AI Cannot Deliver on Its Disruption Promise
The core narrative sold to investors, according to Doctorow, is that AI will 'disrupt' labour markets by replacing human workers, allowing bosses to pocket half the saved salary and pay the rest to AI firms—a story he says Morgan Stanley values at $13 trillion.
However, Doctorow contends this is a fantasy. AI cannot do your job; it can only assist. In fields like radiology, AI might help spot tumours, but the business model is to fire radiologists and have a remaining skeleton crew oversee AI outputs, becoming 'accountability sinks' who take the blame for AI errors.
This flaw is stark in software development. While AI can generate code, it often 'hallucinates' plausible-but-fake software libraries. Malicious actors can then create libraries with those exact names to inject malware. The tedious job of spotting these subtle bugs falls to the 'reverse centaur' coder. Crucially, the experienced, high-wage senior developers who might catch these errors are precisely those bosses aim to replace.
Similarly, AI art's purpose, Doctorow says, is primarily marketing. The total wages of commercial illustrators are minuscule compared to AI development costs. The hype around replacing artists is designed to convince the public of AI's transformative power, despite the 'eerie', hollow nature of most AI-generated content.
Fighting Back: Copyright and Sectoral Bargaining
Doctorow dismisses expanding copyright law to cover AI training as a solution, arguing it would only benefit media conglomerates, not individual artists. Instead, he highlights a powerful existing tool: the US Copyright Office's position that AI-generated works cannot be copyrighted. This means any AI output by a corporation enters the public domain, a major disincentive for its commercial use.
The most effective defence, he proposes, is organised labour. He points to the success of the Hollywood writers' strike, enabled by sectoral bargaining—where all workers in a sector negotiate with all employers. This practice, outlawed for most US workers by the 1940s Taft-Hartley Act, gave writers decisive power.
The Salvage from the Wreckage
Doctorow predicts the AI bubble will burst, with most companies failing and datacenters shuttered. The salvageable remnants will be useful, narrow applications—like transcription or image editing tools—that run on cheap, post-bubble hardware. 'We would call them 'plugins',' he notes.
The fight, therefore, must target the bubble's roots: the myth of job replacement, the monopolists' need for growth narratives, and the false division between workers and the public. 'A serious fight against AI strikes at its roots,' Doctorow concludes, 'the material factors fueling the hundreds of billions in wasted capital.'
