Peter Königs

“There’s this new AI tool, and it’s awesome” doesn’t make for an acceptable AI ethics paper — but “There’s this new AI tool, and it’s deeply problematic” does. As a consequence, the field of AI ethics suffers from excessive negativity about AI. link.springer.com/article/10.100…

@chiaracordelli @Philos_Beard @felix_hohlfeld @StehrPhilipp @SpiekermannKai philevents.org/event/show/139…

Join us at TU Dortmund on June 25-26 for the 2nd Dortmund Disputes on Economic Justice. 🔹 Keynotes: @chiaracordelli, & Bas van der Vossen 🔹 Further speakers: @Philos_Beard, @felix_hohlfeld @StehrPhilipp, @SpiekermannKai, & more

@Phil_Engineer Thanks for sharing your thoughts. Will check out your piece on space ethics after the break. The cultural disconnect is definitely also an issue (should have mentioned it in the paper).

Neat paper! Hope you got good feedback from Cowen’s highlight I’ve worked a bit on this in ‘space ethics.’ I think you well highlight disciplinary norms on content, negativity, incentives, but I think your argument could be made deeper by looking at broader cultural disconnects of humanities from tech/engineering developers. In writing below, I did worry about the types of biases you note and tried to make the report be framed from outside a singular disciplinary frame. Ref, especially observation 2 on cultural differences between engineers and humanities scholars: ntrs.nasa.gov/citations/2023… -Zach

“There’s this new AI tool, and it’s awesome” doesn’t make for an acceptable AI ethics paper — but “There’s this new AI tool, and it’s deeply problematic” does. As a consequence, the field of AI ethics suffers from excessive negativity about AI. link.springer.com/article/10.100…

@pizz4lov3r @torte81 I really like this bird identification AI @MerlinBirdID

Join us for a workshop on 'Autonomy & Algorithms' in Karlsruhe, featuring @carinaprunkl, Simona Chiodo, Keith Harris, Nicola Mößner, and Otto Sahlgren (hosted as part of the Ethics of State Mass Surveillance DFG project). philevents.org/event/show/143…

We are happy to announce the 2nd Dortmund Disputes on Economic Justice, taking place on June 25–26, 2026, at TU Dortmund. The keynote speakers will be @chiaracordelli and Bas van der Vossen. Deadline for abstracts: February 1, 2026. philevents.org/event/show/139…

@MelkumyanK Interesting. Will try out!

@Peter__Koenigs Really interesting paper — the “negative bias” dynamic you describe in AI ethics is spot-on and lines up with what I’ve seen in real reasoning traces as well. I’ve actually built a tool that makes those hidden priors and asymmetries visible in real time: the Epistemic Motion Engine (eme.eagma.com). It shows how conclusions shift when you tweak evidence weights or normative assumptions — like watching publication-bias effects unfold live. If you’re curious, feel free to try it out — or even better, send me a prompt you’d like to see the system think through, and I’ll share the full reasoning session with you.

@AStrasser116 @StefanFSchubert @robsica I mostly agree. Your point that 'good new tech does not put any demands on us' is also made in the paper. I think this is part of the explanation of the paucity of positive pieces on tech. They would have few practical implications.

@Peter__Koenigs @StefanFSchubert @robsica the rise of new technologies like chatgpt? I think that's plausible. Now maybe it's still a problem that the publication landscape is skewed even if most people recognize the "correct answer", but i think it's only a real problem if that led to the view being much more widespread

Paper arguing that ethics papers have an in-built negativity bias. "The negativity crisis of AI ethics", by @Peter__Koenigs, via @robsica link.springer.com/article/10.100…

@AStrasser116 @StefanFSchubert @robsica Any examples? There might exist some, but probably not many. Also, many 'positive' papers are basically responses to others saying 'x is bad/should be banned'.

@StefanFSchubert @Peter__Koenigs @robsica Idk it kinda seems like papers do exist arguing that there has been progress in x, or that x new technology is good and even should be used (re: IVF and lots of adjacent debates). I suspect any negativity bias is in virtue of asymmetric weighting between neg. and pos. generally

@davidmanheim Strongly angree! At least within academic AI ethics, this may have to do with professional incentives. You must criticize in order to publish and have a career. (I'm working on a paper on this.)

The focus on ethics for AI reminds me very much of discussions in bioethics, where far too much discussion is on sins of commission. For example, why not discuss when the bias of AI systems is less severe than that of humans, and *not using AI* should be ethically unacceptable?

Join us at TU Dortmund on June 12-13 for the 1st Dortmund Disputes on Economic Justice. 🔹 Keynotes: @mirandaperrygrl & Nicholas Vrousalis 🔹 Further speakers: @EtyeSteinberg, @kikseven2, @StehrPhilipp, @erik_malmqvist & more ipp.ht.tu-dortmund.de/details/12-130…

Call for Papers! The 2nd Dortmund Conference on Philosophy & Society (Oct 1–2, 2025) features Kate Vredenburgh (LSE) on AI & the future of work. Submit your paper by May 31! ipp.ht.tu-dortmund.de/institut/aktue…

Excited to be teaching a class on 'Libertarianism and its Critics' next term, featuring readings by @bryan_caplan @johanknorberg @mirandaperrygrl @Mattzwolinski @cafreiman @DanDemetriou @landemore @IngridRobeyns

@reidhoffman Thought this might be interesting - resonates well with your Superagency (which I loved)!

Check out my defense of 'surveillance capitalism' link.springer.com/article/10.100…

We are hiring! Apply for one of three PhD positions in a DFG-funded project on the ethics of state mass surveillance (2 at TU Dortmund, 1 at KIT Karlsruhe). Deadline: January 15. German language skills helpful but not mandatory. tinyurl.com/dortmund-docto…

More information about the project at: compphil2mmae.github.io/research/esms/. See also the job ad for the position at KIT: pse.kit.edu/karriere/jobof…. Applications will be pooled. If you apply for one, you'll automatically be considered for the other. No need to apply twice.