From binary logic to Fuzzy Logic
The specialist of Post-Science in Fuzzy Logic Hugh Ching wrote an article(14) about a method(15) to offer quantitative solutions to problems which will be reach in 2100-2500 («the solution of value will replace morality and religion»), the current scientific method is empirical verification. And in around two or three millennium, the method and so AI will be able to do everything the creator did, “the Robot will become the human, and software, DNA”. He explains that fuzzy logic improves methods because “reality is fuzzy” (“precision is sacrificed during the process of expanding a creation’s range of tolerance”). “The range of tolerance of the living system must be wide enough to cover all of the possibilities of a permanently uncertain future”.
What are AI challenges ?
The Dutch Policy Adviser in Public Affairs Bernard said « One challenge will be to use robots and AI to supply the 9 billion people in 2050 with water, food, housing, schools & trainig and so on ». And The Astrophysicist Stephen Hawking believe that this task isn’t sufficient: in 2100 we will be obliged to send people on Mars or elsewhere to survive.
An American computer science student Jack Bandy wrote an article(16) about the consequences of using new technology and he was worried about “framing automation as an economic dilemma rather than an ethical dilemma”. He gives concrete examples of substitution of men by machines not only in the manual sectors but now in the creative and interesting jobs (architects, airplane pilots, ..). The AI & Security Threats report(17) explains that “AI systems can exceed human capabilities” (given the example of chess) in many ways but “they can also fail in ways human never would” (for example data poisoning attacks). Disinformation (political campaigns,..), criminal cyber-offenses (terrorist attacks,..). They explain that “attackers are likely to leverage the growing capabilities of reinforcement learning, including deep reinforcement learning”, some “laws and norms proscribe certain actions in cyberspace but this legal enforcement is difficult across national boundaries”. And their conclusion is that it must be made a priority and analyzed AI from this security angle at all levels.
So, it is urgent to set up a new economic and social model to integrate all of these concerns...