This is quite a topic to ponder about but I am particularly interested in knowing the strategic placement of human-centered inquiries in the advancement of contemporary AI systems despite the remarkable progressions in machine learning and particularly developments in the computer vision. How should we label the fusion, if any, and explicate various intriguing topics of discourse for the mutually beneficial association of two (ostensibly) mechanistically distinct intelligences. Why Human-inspired or Human-assisted AI might be important? Do we really need this association or let machines figure out their own way somehow through continious exploration? What makes humans special and/or not so special? Maybe as humans we have biases towards our species. For one thing, humans are not expert in “detection” missions. Particularly in absence of “attention”, their performance exponentially degrades leaving no room to excel in any task completion. Humans are not good in resolving “hidden correlation” tasks either. Their ability to estimate the joint probability distribution of a very large set of random events is extremely limited. However, they are good at trying/searching and finding out things, in other words, quite good at ontologies. Since the current AI research is oriented around reaching out the level of human-AI, one of the essential questions could be "What is missing to achieve Human-level AI?". An argument that is given is this: Without abstraction, reasoning, compositionality and fuctuality, it is almost impossible to achieve or get to human-level AI in the near future. Of course such a question brings out other interesting and relevant questions such as "What are the most obscure unsolved problems in the development of human-level AI?" or "Can human-specific attributes (coinciousness, morality, humor, causality, trust) can be transfered to machines? If yes, how and to what extent?". Even if we hypothetically assume human-inspired AI is in place, we still need to answers questions like "would this system be fully automatic or partially automatic and Human intervenable?". Should human counciousness position itself somewhere? if yes where in the generic design?. I think for the futuristic designs of AI, we need to look into the possibility of hybrid designs and think about their implementational details along the way...
Dear Suayb, dear participants,
the subject of this discussion is very intriguing for me, and I want to give a first contribution.
I think that the present time debate on AI, with the worries and fears it is raising, is biased by a fundamental misunderstanding. Humans and AI machines are usually compared as if they were easily comparable; the hidden idea is that AI machines and human brain do work in a same way. What if this identity in their functioning were not feasible?
The computer metaphor to understand and explain the brain is a proposal of cognitivism, a set of theories which exactly interpret brain functioning as a series of information processing and calculations. Such vision dominated in all the second part of XIX Century; however, this is the same period in which new theories emerged, in particular those collected under the label "embodiment".
The "embodied" theories have a radically different vision about which, for the moment in this reply, I restrict my contribution to what follows: they (in several different ways) attribute a fundamental role in cognitive processes and understanding to the body. It could be not just a question of brain, but, simultaneously and jointly, of the entire human organism.
So, and it is not a joke, what is missing in AI machines to achieve fully human standards could just be a human body.
I have more exhaustively treated these questions in several works; pardon me for self-referencing: file:///C:/Users/ARPA-Firenze/Downloads/Between_instincts_and_reason_understandi.pdf ; Article Contributions to a neurophysiology of meaning: The interpret...
Best regards
Roberto
Good psychology research is VERY rare (I would even say : non-existent). If proper and real science develops truly in psychology and such is done, it will be very helpful, CRITICALITY HELPFUL.
- Badges
- Science topic
- Similar topics
- Geoscience
- Asia
- India