Generative AI: a neo-cybernetic analysis

PhD ceremony: Mr. C.M. van der Zant, 16.15 uur, Academiegebouw, Broerstraat 5, Groningen

Thesis: Generative AI: a neo-cybernetic analysis

Promotor(s): prof. L.R.B. Schomaker

Faculty: Mathematics and Natural Sciences

The intelligence of machines is still created by humans. As long as machines are not capable of developing their own intelligence, they are merely displaying the cleverness of the designers. The involvement of humans for the scaffolding of subjectiveness in machines hampers the development and autonomy of machine intelligence in profound ways. The mind-body problem, addressed by any researchers in the robotics community, should be re-stated as the mind-body-scaffolding problem, but no general theory of this scaffolding principle exists. If machines are to scaffold their own minds, then they require generative mechanisms to construct and test their mental machinery, not unlike animals and humans. This thesis is about the generic mechanisms behind the automated scaffolding of the intelligence of machines. The developed theory, called Generative Artificial Intelligence, is based on modern physics, biology, sociology, robotics, artificial intelligence and neuro-psychology. Using this new perspective previous research is revisited in order to study which components already for the creation of a generative intelligent machine. For the development of this theory, two notoriously difficult technological problems will be explored, i.e., the automatic recognition of historical cursive script as well as the development of home-robotics applications.