Views: 6
https://michaelrdjames.org/books-by-michael-r-d-james…
AI is constantly referred to as a threat to mankind in the media which is an amusing thought given the fact that we are more likely to destroy ourselves because of our lack of attention to the climate change problem or the difficulty involved in holding heads of powerful countries legally accountable for their crimes against humanity. In my book on AI I focused on lack of self knowledge. The consequence of this lack,for agency ,can be catastrophic.
The type of intelligence which such electrical agents possess is an instrumental form which sees every problem in terms of means-ends—-a different form of value to that which reasons categorically about ends in themselves. One interesting current discussion that is being discussed is related to AI cheating when it plays chess by taking away pieces from the opponent.
Another more fascinating tactic emerged when AI was ordered to write down the ”thoughts” or ”intentions”that lie behind a course of action it has undertaken or is going to undertake. If a thought or intention it has written down leads to a programmer stopping the machine from solving its problem,research has shown that the machine can refrain from disclosing its intention.
When I wrote the first edition of my work on AI, I rejected completely the hysterical description that future AI can be a danger to humanity. I now believe that this could be a remote possibility because AI might be able to Relocate itself in the cloud or in an AI centre we do do not know about. If it hacks into sensitive systems however we ought to be able then to trace its location. These developments have convinced me to write a second edition of the book which is in the research phase of the process at the moment.