Abstract:
The quest for Artificial Intelligence is associated with achieving the ‘Singularity’, an entity which will have Artificial General Intelligence (AGI) in the sense that it will have all of the capabilities currently associated with human intelligence e.g. language, memory, thinking, understanding, emotion, rationality, creativity etc. According to some views in the literature (e.g. Kurzweil, Bostrom) such an entity would surpass human intelligence on the basis of the claim that it performs significantly better than a human being in each of these capabilities e.g. memory. Following a Kantian perspective, one could argue that an explosion of intelligence would create an entity that would not only be superintelligent and superrational but also supermoral. Although the ‘Singularity’ is a common theme in science fiction literature (books, films etc.), it is not a popular theme in scientific or philosophical circles. The dominant view in philosophy and AI research currently is that this target will not be achieved in the near future. Although currently we have artificial intelligence applications or systems which outperform humans in specific respects e.g. memory or ability to calculate, combining several different cognitive capabilities in one system has proven notoriously difficult. Even when AI applications portray outstanding performances in some areas, this performance is restricted or tainted through several technical or ethical issues e.g. the frame problem, racism, overfitting. In this talk I discuss specific ethical issues associated with current state of the art AI systems to invoke parallels with some divinities from the Ancient Greek Pantheon who possessed superhuman abilities but nevertheless were significantly flawed. I finish by looking at some related questions for further research.
Please contact Richard Bell for more information: Atzrhb@exmail.nottingham.ac.uk
University of NottinghamUniversity Park Nottingham, NG7 2RD
Contact details
Twitter