What exactly is strong Artificial Intelligence (AI)? Strong AI is a term used to define a form of artificial intelligence comparable to that of a human. If such an intelligence were to exist, we would not need to cram gigabytes of examples into computers in order for them to solve complex problems. Moreover, the existence of strong AI would imply a machine capable of deliberate and conscious reasoning, as well as feeling emotions and expressing intentionality and creativity. When will this kind of AI exist? Not tomorrow, that’s for sure. There are those who strongly believe that human-level general intelligence is possible, such as Ray Kurzweil, a transhumanists and Google’s Director of Engineering. While the feat accomplished by the DeepMind’s AI-based Alphazero is impressive (it managed to beat Stockfish, the best chess program, after only a couple of hours of training or learning), this remains a form of weak AI. It simply demonstrates a significant jump in a computer’s ability to learn and deal with combinatorial explosion in the closed world of chess. It has only one objective: to kill. We should remember that the expression check mate comes from the Arab expression al cheik mat, meaning the king is dead.
Mozart teaches us that music and musicians are timeless. This is true in other domains of human intelligence such as the arts, philosophy, ethics, love, etc. These are the abilities that define our lives and give them meaning. That which gets us going, but remains a choice. That which transcends us. A human made up of models and experiences (his/her datasets) would be one-dimensional. We would consider such a person unintelligent..and dangerous.
Spider-man’s uncle said “with great power comes great responsibility”. It is understandable why Elon Musk has warned us about the dangers of using AI for military objectives. His arguments bring forth important ethical questions and convey the need for world-wide regulation on this subject. For me, as the CEO of a technology company and an active participant in this technological revolution I am more concerned with delegating out intelligence to a weak AI, than the emergence of a strong AI that can rival the intelligence of humans. One needs to be realistic; before the arrival of Terminator there are the Trumps of this world. Weakness can emerge from things other than weak AI.
It is easy for decision-makers to be seduced by the spectacular feats accomplished so far by current AI systems. To give into our desire for intellectual comfort and to delegate our conscience to these automated chess-players. There is a need for ethical regulations because weak AI should not be synonymous with comfort or a decline in the use of human qualities for our strategic decisions. These regulations are needed to question and perfect weak AI, in order to make both human and artificial intelligence more intelligent.
Written by Bernard EUVERTE, CEO&founder of WorkIT Software