AI Now Institute: Doomerism Is Science Fiction That Favors Big AI
AI Now Institute argues that fears of AI-driven human extinction rest on no scientific work and end up serving the largest AI companies. In an article published on its site, senior research scientist Boyan Milanov calls the claims "science fiction."
- AI Now Institute says there is no clear path from today's AI systems to superintelligence.
- Senior Research Scientist Boyan Milanov calls AI doomerism "science fiction" that is not scientifically verifiable or falsifiable.
- The article says leaders of companies like Anthropic and OpenAI claim only their organizations can build AI responsibly.
- Milanov describes this as self-regulation rather than audits by independent parties acting in the public interest.
No clear path to superintelligence
The core argument of the piece is that there is no clear path showing that today's AI systems will lead to superintelligence. According to the article, there is also no basis for calculating the probability of doom, a figure that is often cited in debates about the long-term risks of advanced AI.
The article notes that the power attributed to superintelligence has led some in these communities to describe advanced AI in God-like terms. Milanov, a Senior Research Scientist at AI Now Institute, puts it more bluntly: "They believe in the advent of an almighty powerful AI god, which will either lead to destruction of humanity or help us achieve utopia."
"Science fiction," not science
Milanov rejects the idea that doomerism has a research foundation. "What we have to keep in mind is that all this doomerism isn't based on any scientific work," he says. "It really is science fiction. None of the claims are scientifically verifiable or falsifiable."
The distinction matters because claims that cannot be tested cannot be checked by outside researchers or regulators. That is the point the article builds on when it turns to the companies themselves.
Self-regulation and who benefits
The article says the heads of major AI companies such as Anthropic and OpenAI each indicate that their organizations are the only ones who can build these systems responsibly, with guardrails that the companies themselves will decide on.
According to Milanov, this approach amounts to self-regulation. Companies would set their own standards and processes, instead of being audited by independent parties who work in the interest of the public.
He applies the criticism to a specific case. "If you read between the lines, everything he advocates [for] is something that would help companies like Anthropic," he says. The excerpt does not name the person Milanov refers to.
What it means
The piece frames the doom narrative as more than a scientific disagreement. In AI Now Institute's view, the framing can strengthen the case for letting leading developers define their own safety rules, rather than submitting to independent audits. The excerpt published on the institute's site is short and points readers to the full article for the complete argument.
AI Now Institute: How the bad science of AI doomerism is good for big business
