On 5/1/24 12:38 pm, David wrote:
At the following link, I run the argument that 'AI' was great as a means of getting a research grant in 1965, but the misconceptions embedded in it have done a lot of damage already and will do a lot more shortly:On 03/01/2024 19:49, Narelle Clark wrote:From the November issue of the IEEE Communications Magazine there are some papers on 'Semantic Communication': "In contrast to the Shannon paradigm that focuses on correct reception of the transmitted packet regardless of its meaning, semantic communication is concerned with the issue of how to efficiently transmit and receive the desired meaning of the source content to the destination. By transmitting only the meaning or semantics of the source content, semantic communication holds the promise of making wireless networks significantly more energy-efficient, robust, and sustainable than ever before."Just what we need embedded in everything, huh? AI determining the "meaning" and summarising the lower layer communications for more efficient transmission.I can think of an even more efficient way of communicating: forget about all those difficult theoretical concepts like Shannon's Equation or Entropy, and let the listener just assume they know what the human at the other end is saying! Why bother about physics, it's all too difficult and too hard to market.I must be feeling in a cynical mood this morning."Semantic communication" is quite often known as poetry, which conveys its message by conjuring the listener's emotions and imagination, their shared experience of life, and their empathy for others.However humans have been developing such understanding and mental capacity one way and another since the dawn of sentient life, and the human brain began its evolution perhaps two million years ago. Furthermore, human cerebral reactions cannot be separated from the rest of the body's biochemistry, such as the endocrine system, the sympathetic & para-sympathetic nervous system, and so on.This is essentially why AI machines should be conceived (by humans!) as fast correlation processors. That's not to deny they're an extremely important technology with vast potential to improve the lot of humanity, and also to destroy it. I don't agree that AI is some sort of technology bubble which will just disappear given time.Yes? No?
http://www.rogerclarke.com/EC/AITS.html#RAIIf, and only if, we re-conceive and re-orient can we can get the threats back under control, and reap benefits.
What's needed is not 'AI', but 'Complementary Artefact Intelligence'.And that requires the use of decision-support thinking, envisioning Artefact Intelligence as being *designed to* integrate with Human Intelligence, to produce Augmented Intelligence.
And, while we're at it, we need to build an explicit linkage with robotics - or better still with 'co-botics' - and talk about complementary artefact capability combining with human capability to deliver augmented capability:
http://www.rogerclarke.com/EC/AITS.html#F2 -- Roger Clarke mailto:[email protected] T: +61 2 6288 6916 http://www.xamax.com.au http://www.rogerclarke.comXamax Consultancy Pty Ltd 78 Sidaway St, Chapman ACT 2611 AUSTRALIA
Visiting Professor in the Faculty of Law University of N.S.W. Visiting Professor in Computer Science Australian National University
OpenPGP_signature
Description: OpenPGP digital signature
_______________________________________________ Link mailing list [email protected] https://mailman.anu.edu.au/mailman/listinfo/link
