(in Polish) Religion, Logic and Artificial Intelligence

elective monographs
general courses

Can chatbots talk about religion in a way that it really makes sense? Are they helpful in the logical analysis of religious arguments and religious discourse? What about their own discourse on religious issues?

The course “Religion, Logic and Artificial Intelligence” is aimed at providing participants with basic knowledge about tools useful for analyzing religious discourse and about machine learning and training AI models.

The main goal of the course is making participants ready to: assess effectiveness and possible use of chatbots in the analysis of religious discourse, and analyze chatbots’ statements on religious topics.

The course will be unified by philosophical questions about: the development of AI and its rationality, the role of logic in the context of AI, the nature of thinking, the status of discourse and the future of religious discourse.

Can chatbots talk about religion in a way that it really makes sense? Are they helpful in the logical analysis of religious arguments and religious discourse? What about their own discourse on religious issues?

Chatbots based on Artificial Intelligence (AI) models have become important discourse actors. They generate content on portals, they take part in discussions, they are our partners for conversations. The aim of the course will be to collect and test tools for assessing the potential of chatbots as rational actors in one of the most important (for its impact on human behavior) part of the discourse, which is the religious discourse.

To achieve this goal, participants of the course “Religion, Logic and Artificial Intelligence” will learn about various tools useful for analyzing religious discourse (the latest results of research on the relationship between logic and religion will be taken into account), then they will gather basic knowledge about machine learning and training AI models. On this basis, they will be able to assess the effectiveness and possible use of chatbots in the analysis of religious discourse (including arguments), and then to analyze chatbots’ statements on religious topics. The unifying theme of the entire project will be philosophical questions about: the development of AI and its rationality, the role of logic in the context of AI, the nature of thinking, the status of discourse and the future of religious discourse.

In order to get diverse and up-to-date knowledge, guest participation of experts in some meetings is assumed.

The following topics will be discussed during the lecture:

1. Argumentation in religious discourse: definitions and an overview.

2. Argument annotation and diagramming (including Inference Anchoring Theory) applied to religious discourse.

3. Logical problems in religious discourse, and logical solutions.

4. Logical interpretation of religious discourse with the use of: paraconsistent, non-monotonic and fuzzy logics.

5. Logical approach to the interreligious dialogue.

6. Computer analysis of religious arguments.

7. Chatbots based on Artificial Intelligence and the nature of their “reasoning” (the idea of machine learning and philosophical inspirations: Wittgenstein, Turing, Foucault).

8. AI-chatbots as rational actors in religious discourse and interreligious dialogue (“skills audit”).

9. AI-chatbots as a help in the analysis of religious arguments.

10. Logical assessment of AI-chatbots religious discourse.

11. Philosophical questions about AI-development, its rationality, and the future of religious discourse.

Active and creative participation of the participants is welcome.

- Annotation Guidelines for Inference Anchoring Theory (IAT) with support for Conventional Implicatures (CIs), Centre for Argument Technology, www.arg.tech, April 2018, https://typo.uni-konstanz.de/add-up/wp-content/uploads/2018/04/IAT-CI-Guidelines.pdf

- P. Balcerowicz, Logic in religious and non-religious belief systems. Int J Philos Relig 84, 113–129 (2018). https://doi.org/10.1007/s11153-017-9646-x

- E.D. Bohn, E.D. The Logic of the Trinity. SOPHIA 50, 363–374 (2011). https://doi.org/10.1007/s11841-011-0265-1

- T. B. Brown et al., Language Models are Few-Shot Learners, arXiv:2005.14165v4 [cs.CL] 22 Jul 2020 (fragments).

- A. Banino et al. Vector-based navigation using grid-like representations in artificial agents. “Nature” 2018, nr 557, ss. 429–433, https://doi.org/10.1038/s41586-018-0102-6

- G. Priest, An Introduction to Non-Classical Logic: From If to Is. New York: Cambridge University Press 2008.

- Yu Sun et al., ERNIE 2.0: A Continual Pre-Training Framework for Language Understanding, arXiv:1907.12412v2, 21.11.2019.

- N. Slonim et al., An autonomous debating system, “Nature” 2021, nr 591, ss. 379–384, https://doi.org/10.1038/s41586-021-03215-w

- knows the basics of argumentation theory;

- knows examples of logical problems in religious discourse;

- knows basic methodology of testing AI chatbots' abilities in the selected scope (research concept, tool, tool test, implementation);

- knows the basics of training natural language models;

- knows basic differences between natural language models.

- applies the basics of argumentation theory to discourse analysis;

- applies basic knowledge of logic (including non-classical logics) to discourse analysis;

- identifies logical problems in the field of religious discourse;

- formulates philosophical questions about AI-based chatbots as actors in religious discourse.

- takes part in philosophical discussions.

The assessment of participants’ learning outcomes will be based on (alternatively, not exclusively):

- presentation of participants’ own projects (concerning testing chatbots’ skills with respect to religious discourse and logic), OR

- handwritten short essays or other kinds of written works, OR

- oral presentation of questions, arguments and answers to counterarguments, OR

- written test, OR

- oral test.

Number of absences: 2