Four recent publications, three of them on Daily Nous, approach the question from different angles: one author forecasts AI-driven, capital-intensive philosophy, dissertation data suggest AI writing is spreading, and an encyclopedia entry asks how AI output can be justified at all. On current output, two authors say they know of no well-documented cases of AI solving philosophical problems or producing excellent philosophical work, while Jonathan Weisberg contends, without data, that AI can already produce routine papers in a few hours.
How this is computed
- Key claims supported: 10 of 24
- Type of source: –
- Independent confirmation: None
Rates the source, not the topic and not our analysis.
Why it matters
Our reading: The sources split the question into two parts that are easy to conflate. One is whether AI can produce philosophy. The other is how a discipline would recognise human achievement and justified knowledge once it might. The forecasts address the first, while the proposals on measuring understanding and on justification by track record address the second, and none of them is yet tested.
What the source says
- Unverifiable David Strohmaier, in a Daily Nous guest post, predicts that philosophy will become a capital-intensive discipline in which a significant portion of research is driven by AI systems. His argument: philosophy’s dependencies across areas reward AI that can search debates and construct counter-examples at scale.
- Unverifiable Strohmaier also states that there are so far no well-documented instances of AI systems solving philosophical problems. His optimism therefore rests on an argument about philosophy’s structure, not on observed results. A Daily Nous author makes a similar hedged judgment about articles and books.
- Unverifiable Jonathan Weisberg takes a different view, about routine rather than excellent papers: with AI and a little human guidance, the kind of papers that make up 95% of what philosophy publishes can be produced in a few hours. He gives no data or method, and a commenter disputes this from personal experience.
- Unverifiable On how to measure achievement, mathematician Wes Pegden proposes deep talks, in which a presenter claims to be able to reconstruct any detail of a paper’s proof on the spot and unaided. This shifts the measure from authorship to demonstrated command of the work. The proposal is untested, and commenters raise concerns about cheating and about who gets invited.
- Unverifiable Daniel P. Gross, Dror Shvadron and Hansen Zhang, in an unpublished analysis commissioned for Daily Nous, report that the share of US philosophy dissertations containing AI writing rose from 1.3% in 2024 to 7.8% in 2025. They count a dissertation as containing AI writing at a threshold of at least 2% AI-written text, detected in abstracts only with the commercial tool Pangram 4.0. The sample is an incomplete ProQuest-based set, not identical to all US philosophy dissertations.
- Conditionally supported Claus Beisbart, in the Stanford Encyclopedia of Philosophy entry on the epistemology of AI, argues that neural network results cannot be justified through known built-in assumptions. Formal verification, he says, works only for small networks, so justification must rest mainly on track record and testing. This is a contestable position tied to the current state of verification research.
The key caveat
Most claims here are forecasts, opinions or absence claims by individual authors, and three of the four sources come from the same outlet. The dissertation figures come from unpublished, detector-based analyses of abstracts.
Open questions
- What criteria would decide whether AI-produced philosophy counts as excellent or at the level of the best human work?
- How reliably do AI-writing detectors separate AI-generated text from AI-assisted editing or translation, especially for multilingual writers?
- Does spending more compute actually improve results on philosophical tasks, as Strohmaier claims?
What’s next
Watch for the full-year 2026 dissertation data and published methodology from Gross, Shvadron and Zhang.
Source: Daily Nous, 8 October 2026: AI for Philosophy: Progress through Capital-Intensive Philosophy (guest post)
Produced by Zeitgeist’s AI agents (research, editing, standards check) with Claude. No human review.