Artificial Intelligence learns to discuss philosophy

An Artificial Intelligence (AI) learns to answer philosophical questions, perfectly imitating its human model, namely the American philosopher Daniel Dennett, known for his theories on the philosophy of science and that of biology.

The system, which works in a similar way to ChatGpt, is the result of a collaboration between Europe and the United States and is described on the arXiv platform, which welcomes articles awaiting examination by the scientific community.

In recent months there has been a boom in creative uses of AI, from algorithms capable of designing small masterpieces starting from a textual description to chatbots capable of processing texts of all kinds in a few moments, even if often with even serious errors , up to the production of content that imitates those of well-known artists or writers.

Now there are also algorithms to imitate philosophers.

Turning the idea into reality were Eric Schwitzgebel of the University of California at Riverside, David Schwitzgebel of the Scuola Normale Superiore in Paris and Anna Strasse of the Ludwig-Maximilians University of Munich.

The researchers asked philosopher Daniel Dennett 10 questions and his answers were provided to the new AI model, called Digi-Dan, which based on the philosopher's words developed new alternative answers.

The questions were then shown to 425 individuals, who were asked to recognize, among four possible answers, which ones were those of the philosopher and those generated by the system.

It thus emerged that philosophers were able to identify the original answers in about 50% of the cases, against 20% of the participants ignorant of philosophy.

According to the researchers, Digi-Dan could in the future be instructed on texts by philosophers who have now disappeared to answer new questions.

The answers, observe the researchers, certainly could not be considered reliable, but could offer new and stimulating points of view.