Artificial Intelligence, Freud, and Asimov

Can Artificial Intelligence (AI) ever capture—and not be captured by—the unconscious? So far, few authors address this possibility.1 AI's effects on consciousness, on the other hand, yield a vast outpour of critical reflections. AI's goal to attain, or at least persuasively simulate, human consciousness preoccupies, for good reason, contemporary debate. What I propose instead here is to look at AI's ultimate goals in light of basic Freudian concepts and so highlight the narrowness of the conventional debate. Thus, we invoke Freud's model of the mind as three compartments—conscious, preconscious and unconscious— and how it illuminates crucial barriers in achieving non-human consciousness. We then do the same with the classic tripartite of id, ego and super-ego. This modest exercise produces interesting fare for further thought.

This article will not only treat how psychoanalysis views consciousness in contrast to the philosophers and neuroscientists but also argues the need for vaunted AI to incorporate unconsciousness in order to encompass the human mind in the most (artificially) complete sense. Why try to recreate these mental processes in AI? Because it might be the only way for humans to ensure and/or maximize safety against AI as we move forward with this huge keg of dynamite.

Science fiction frequently is precursor to scientific reality. So as an illustration we turn to science fiction for a situation that illustrates AI's apocalyptical potential: Isaac Asimov's I, Robot. Asimov's book explores the gradual takeover by robots notwithstanding humans inputting three safety rules programmed within robot, ahem, DNA. These rules ensure that robots never harm or depose humans. So we will compare what is conjectured by Asimov versus likely reality if we follow the current AI path, and what dangers could be avoided if key Freudian concepts are included in discussions and in AI models too.

Has AI Achieved Consciousness Already?

For the most part, AI experts, alb