Intentional and Phenomenological Aspects of Consciousness in Human-Computer Interaction

 

Presented at the 5th Graz  Congress on Medical Hypnosis 2020

Author. Paul Seitlinger

 

Affiliations. Tallinn University (School of Educational Sciences, Estonia), University of Graz (Institute of Psychology, Austria)

Abstract. New opportunities in representing and analyzing our traces in the digitized world have spawned progress in building models of the human mind. These developments are fueling heated debates in cultural, social and educational sciences as well as research on human health.

Typically, these debates boil down to the question: Are emerging applications of artificial intelligence (AI) increasingly conscious of what they are doing? If the answer were yes, it would probably lead us to spectacular human-computer interactions as envisioned by more recent science fiction stories. In the movie “Her” [1], for instance, a newly developed algorithm is not only capable of understanding a user’s feelings but also of generating and communicating its own emotional states. Together with its superhuman efficiency in making sense of what the user is saying, the algorithm soon becomes an irreplaceably empathic partner and even object of desire. 

The goal of this talk is not to pass any moral judgment on such fiction but to demonstrate the huge gap that parts our present technologies from a “Her”-like algorithm. To this end, I will start with a brief sketch of two facets of human consciousness: the intentional and the phenomenological, with the latter being essential for the awareness and communication of subjective experiences (qualia) in an interpersonal dialogue. I will then put forth some well-considered arguments of philosophy of consciousness, according to which present AI applications can indeed be deemed conscious, but primarily in an intentional sense [2]. 

After this theoretical introduction, I will aim to concretize these arguments by some empirical examples taken from my own cognitive-psychological work in the field of Webscience and Technology-Enhanced Learning. I will tell from an interdisciplinary research among people from computer and learning science, in which we are applying state-of-the-art neural networks to model and predict user traces (e.g., verbal and information behavior) in social Internet and learning environments [3, 4]. Though our own modeling attempts and those of related research groups are successful in capturing intentional aspects of a user’s consciousness, an access to their phenomenological experiences seems to lie out of reach.

Towards the end, I will report from some of our more visionary goals in the field of human-computer interaction [5], by which we aim to account for a user’s qualia as well (e.g., valence and intensity of emotional states). Paramount to this endeavor will be the extension of recorded data samples by physiological data that, for example, can be stimulated and measured through vibro-tactile devices. Coupling such recordings to psycho-physical measurement models should contribute to conceptualizing more humanoid algorithms who are simulating and communicating their own qualia. I will close the talk by outlining practical implications of such concepts in settings to be found in the educational and clinical context (e.g., psychotherapeutic sessions).

 

References

[1] Her (2013). Box Office Mojo. Internet Movie Database.

[2] Gabriel, M. (2017). I am not a brain: philosophy of mind for the 21st century. Polity Press.

[3] Seitlinger, P. & Ley, T. (2016). Reconceptualizing imitation in social tagging: a reflective search model of human web interaction. Proceedings of the 8th International ACM Web Science Conference 2016 (pp. 146-155). New York: ACM press.

[4] Seitlinger, P., Ley, T., Kowald, D., Theiler, D., Hasani-Mavriqi, I., Dennerlein, S., Lex, E., & Albert, D. (2018). Balancing the Fluency-Consistency Tradeoff in Collaborative Information Search with a Recommender Approach. International Journal of Human-Computer Interaction, 34 (6), 557-575.

[5] Zhiglova, Y., Lamas, D., Smorgun, I., & Seitlinger, P. (2019). Vibro-tactile implicit interactions: so what? Proceedings of the 17th International Conference on Human-Computer Interaction (in press). Heidelber: Springer. 

Alle Beiträge im Überblick