Cory Doctorow, award‑winning science fiction author, journalist, digital rights activist and special advisor to the Electronic Frontier Foundation (EFF), warns that apocalyptic rhetoric about artificial intelligence functions partly as self‑deception and partly as a business narrative that conceals tangible economic and labour‑market harms. In a comprehensive interview, Doctorow argued the immediate danger from AI is not consciousness, but an unsustainable investment bubble and the irreversible loss of organisational process knowledge.
Two financier types and the promise of the ‘‘humanless company’’
Doctorow divides the elite funding AI into two groups. One is driven by a psychological distortion: wealth and detachment from reality create a preference for a world run by statistical abstractions without humans. The other group views AI purely as a cold business opportunity, recognizing it is easy to sell management a tool that can remove continually dissenting experts. Meanwhile, some researchers inside AI labs genuinely panic and believe their own apocalyptic rhetoric — as Doctorow put it, they are “high on their own supply.”
The endpoint of this logic, he says, is a corporate ecosystem where expensive employees with independent judgment are replaced by obedient statistical systems.
The numbers don’t add up: a trillion dollars of spend versus modest revenues
Doctorow focuses on the economic unsustainability of current generative AI business models. Training and running large models requires vast data centers, specialized chips and significant energy, while subscription fees do not cover these costs. Drawing on Ed Zitron’s analyses, he suggests that free or loss‑leading inference capacity — for example at OpenAI — can function as marketing spend that masks the true economic burden of the service.
He argues the market is fragile, dependent on capital and debt financing, where a larger shock could trigger a domino effect. The article cites a claim that this year roughly a trillion dollars was spent on investments while revenues were about 50 billion dollars.
Wall Street and industry studies also question returns: Sequoia Capital’s widely cited analysis and a report from Goldman Sachs’ global research unit have both raised doubts about whether current infrastructure spending can ever be justified by future revenues, asking whether AI will ever produce the economic returns required to justify the present investment frenzy.
The real damage: loss of organisational process knowledge
One of Doctorow’s most serious economic warnings concerns labour‑market impacts. When companies fire experienced workers and replace them with seemingly cheaper AI, they lose organisational process knowledge — the unwritten, institutional know‑how that often separates profitable firms from unprofitable ones.
That knowledge is hard to restore and its loss can cause generational productivity declines before management recognises that AI cannot fully perform the job. Each time a skilled worker is replaced by an underperforming chatbot, Doctorow says, a piece of organisational memory is destroyed.
AI as the next stage of platform degradation
Doctorow has previously coined the term “enshittification” to describe platform degradation: the gradual decline in service quality as platforms prioritise profit. He argues AI can take that process to a new level. Unlike traditional cases where established profitable platforms degrade quality to squeeze more profit, AI systems are often inherently loss‑making, vast money sinks that must be perpetually sustained.
Technical features of AI also create plausible deniability. Notorious AI “hallucinations” can offer companies a communications and liability escape hatch: if an AI recommends a more expensive product that gives the company a higher commission, the company could dismiss the mistake as a hallucination when the error is exposed.
Conclusion: regulators should focus on present harms rather than speculative apocalypse
Doctorow concludes that AI leaders’ apocalyptic rhetoric is unconvincing if it is not matched by proportionate, substantive safety and regulatory guarantees. Regulators and markets, he argues, should be less preoccupied with sci‑fi end‑of‑the‑world scenarios and concentrate on the everyday economic and labour‑market damages that the pursuit of exaggerated AI promises is already causing.
Selected works and honours of Cory Doctorow
- Little Brother (2008) — John W. Campbell Memorial Award, Prometheus Award, Sunburst Award; nominated for the Hugo Award.
- Down and Out in the Magic Kingdom (2003) — Locus Award (Best First Novel).
- Pirate Cinema (2012) — Prometheus Award.
- Homeland (2013) — Prometheus Award.
- The Internet Con: How to Seize the Means of Computation (2023; Hungarian edition 2026) — Locus Award (Non‑fiction).
(Note: the source article also included promotional mention of a Portfolio AI & Digital Transformation conference scheduled for November 26.)
