Arquivo para October 16th, 2017

(Português) Complexidade, consciência e AI

16 Oct

Complexity, consciousness and AI We have already stated that both AiMachinecomplexity and consciousness are phenomena that belong to the biological nature, but AI (artificial intelligence) continues to have advances, what would be the misconceptions of mistaken notions in this field?

The response of MIT technology professor Rodney Brooks, who works with robotics for Panasonic, is that there are seven deadly sins, and he cites Amara’s law by saying that people tend to underestimate both the short-term and the short-term effect. long-term technology while examining a nonexistent technology, while another is confusing hypotheses where the AI ​​would have an equal competence to solve the problem of a nonexistent technology.

A third factor pointed out by Brooks is that the assumption often of practicing a task is often confused with a task performed by AI equal to competence.

Brooks also says that people are prone to parallel AI progress in learning a given task for the same process in humans, so it always comes up with the idea of ​​human hybrids / machines.

Brooks also asserts that people should not expect AI to continue to make steady progress on an exponential path, but rather on adjustments and reevaluations, and we should not believe in media scenarios with unexpected AI scenarios.

This is the science fiction of Odyssey 2001, where the computer that made diabolic decisions never existed and watch the movie today shows the unreality of that fiction, Blade Runner was updated in 2049 and asks if machines have souls, the writer’s question that inspired the film is whether machines dream of electric sheep (in Philip K. Dick’s novel, and why machines would sleep? and why would machines sleep?

Incidentally the figure of the dog of the old Andromeda songwriter, also calling Harrison Ford for the role was interesting in reference to the novel that inspired the film.

The full article by Rooney Brooks published in Technology Review´s last week is quite interesting and separates 7 fallacies about AI.