LLMs and digital autonomy: from misinformation to context collapse

The conference will be held September 5-7, 2025, at the University of Bucharest, Romania, as part of "The effects of LLM interaction in digital and virtual environments on TOM" ICUB grant in enhancing institutional performance at the University of Bucharest, gathering a research team from Philosophy, Linguistics, Cognitive Psychology and Evolutionary Biology.

It will have a mixed format, in that speakers may choose whether they present online only or face to face at the event's location (if so, their session will enjoy a live audience, but it will also be streamed to remote participants). Regular presentations will be 30 minutes long, followed by 10 minutes long Q&A and 10 minutes breaks.

The conference explores the ways in which LLMs impact the autonomy of users in the digital environment. It targets the issues raised by interacting with such AI, especially when it comes to data privacy of users, data control (the ways in which users get to choose what happens with their data), and the ways in which they influence users' decision-making process.

Panels include, but are not limited to:

1.LLMs, privacy and data control - how LLMs bear on user autonomy.

2.LLMs and their impact on misinformation, echo chambers, fake news and context collapse (opportunities and risks).

3.Accountability, confidentiality, responsibility and intellectual rights attribution when LLMs collaborate with human users (data generation such as text production; therapy bots;

4. Do users and LLMs interact epistemically? Are LLM outputs reliable? What leads to epistemic dependence on LLMs?  Should it be prevented? If so, how?

5. Can we trust therapy bots? (autonomy, trust, and trustworthiness - AI/Chatbot interaction in the context of mental health)

Conference programme

LLMs and digital autonomy: from misinformation to context collapse

5-6th of September, 2025

University of Bucharest, Faculty of Philosophy

Friday, September 5th

Chair : Daniel Stancu (University of Bucharest, Faculty of Philosophy)

14:45 Opening

15:00 - 15:30 Matthew Jope (The University of Aberdeen) - AI assertions, responsibility gaps, and epistemic blame

15:30- 16:00 Mahesh Venkataraman (Accenture) A systems-theoretic approach to mitigating epistemic distortion in LLMs

16:00 – 16:50 Keynote: Zuzanna Rucinska (University of Antwerp, Belgium) Can we trust bots with our mental health? Enactivist and phenomenological perspectives

17:00 – 17:30 – Nic Fisk (University of Rhode Island, Kingston RI, United States of America) and Aarthi Venkat (Broad Institute of MIT of Harvard, Cambridge MA, United States of America) Epistemic responsibility and interdisciplinarity in the age of AI for science

17:30 – 18:00 Break

18:00 – 18:30 Michelle M. Dyke (Tufts University) Do we blame AI? Implications for the debate over epistemic and moral blame

18:30- 19:00  Andrei Mărăşoiu (University of Bucharest, Faculty of Philosophy, CELFIS, ICUB) Automated metaphor comprehension comes natural        

19:00- 19:30 Marc Champagne (Kwantlen Polytechnic University, Canada)How LLMs cause semiotic inflation, which causes experiential devaluation

 Saturday, September 6th

Chair : Elena Roşu (University of Bucharest, Faculty of Philosophy)

10:00 – 10:30 Adam Andreotta (Curtin University) The question is not, can LLMs reason? Nor, can they talk? But, can they contribute to philosophical progress?

10:30- 11:00 Ching Lam Janice LAW (Chinese University of Hong Kong) From infodemic to context collapse: LLMs, misinformation, and the ethics of digital autonomy

11:00 – 11:30 Chiara Russo (Università degli Studi di Catania) What Large Language Models don’t say: bias, tokenism, and discursive constraints in Italian LLMs

11:30 -15 Break

15:00 – 15:30 Adina Barbu (University of Bucharest, Faculty of Psychology and Education Sciences ) Artificial Intelligence, creativity and intellectual rights

15:30- 16:00 Sandra Branzaru (University of Bucharest, Faculty of Philosophy, ICUB, CELFIS) LLM narrative games: writing the "thinking"

16:00- 17:00 Keynote: Ioan Muntean (University of Texas Rio Grande Valley)'Digital phenotyping' and 'precision mental health' with AI: unity, collapse, or pluralism in psychiatry?

17:00 – 17:20 Concluding remarks