rTAIM/LEXISLUR Talk
Norms in Human-AI Assertion
Markus Kneer (University of Graz)
8 July 2025 | 14h30-16h00 | Sala de Reuniões 2
Faculdade de Letras da Universidade do Porto
In this talk, I will discuss several possible accounts of the human norm of assertion: knowledge, justification, truth and belief. On the basis of empirical evidence,
I will suggest that when humans talk to humans, the view that there assertions are governed by a justification norm best captures actual linguistic behavior. Differently put, in order to assert that p, one must assert p for good reasons, though p can be false.
I will then extend the discussion to linguistic human-AI interaction, that is, our normative expectations when it comes to talking to (sophisticated) chatbots. Here, the data will show, a more stringent norm of assertion seems to be in place. I will reflect
more broadly on the difference in norms of assertion across agent-types and trace out its implications for AI alignment in general.
Markus Kneer is professor of Ethics of Artificial Intelligence at the University of Graz, Austria, the director of the Guilty Minds Lab and the leader of the project "Norms of assertion in linguistic human-AI interaction." His work is both theoretical and empirical in nature. Central topics concern mental state and moral culpability attribution, norms of assertion, and a variety of questions in ethics of AI.
Organisers: Dan-Cristian Zeman (2023.05952.CEECIND) & Steven Gouveia (2022.02527.CEECIND)
RG Mind, Language and Action Group (MLAG)
Instituto de Filosofia da Universidade do Porto - UID/00502
Fundação para a Ciência e a Tecnologia (FCT)
____________________________________________
Instituto de Filosofia (UI&D 502)
Faculdade de Letras da Universidade do Porto
Via Panorâmica s/n
4150-564 Porto
Tel. 22 607 71 80
E-mail: ifilo...@letras.up.pt
http://ifilosofia.up.pt/
Financiamento, avaliação e auditoria do Instituto de Filosofia:
