Regensburg 2025 – scientific programme

Parts | Days | Selection | Search | Updates | Downloads | Help

AKPIK: Arbeitskreis Physik, moderne Informationstechnologie und Künstliche Intelligenz

AKPIK 5: Poster

AKPIK 5.15: Poster

Thursday, March 20, 2025, 15:00–16:30, P2

Experimental examination of the validity of the Turing test for considering Artificial Intelligence as having subjectivity — •Lev Gelbart and Alexey Iakovlev — Moscow, Russia

The speed of development of Artificial Intelligence (AI) raises questions about its subjectivity and thinking. Alan Turing, who set the criteria for thinking, did not consider subjectivity. There are sharp discussions on this topic, for example, Doctor of Law Valery Zorkin, professor at MSU, opposes endowing AI with subjectivity, emphasising humanism.

It should also be taken into account that the Turing test may be invalid if we consider people, who are not capable of passing it. For example, a study by Alan Ropper (2010) showed that patients in a vegetative state can demonstrate brain activity, but are clearly unable to pass the Turing test, which does not detract from their human value.

We have considered, whether a machine can pass a test designed for humans. Machines are thought to not be able to adequately mimic emotions, so we tested three chatbots (Lily, AI Chat, ChatGPT) for emotional intelligence using the IDRlabs test. All chatbots performed above the average of humans (77.27%, 68.87%, 64.6%), but this, obviously, does not make them subjects of law.

Conclusion: the Turing test is not valid for assessing the subjectivity of AI.

Keywords: Artificial Intelligence; Turing test; Emotional Intelligence; Subjectivity; Chatbots