Thread: Is AI necessary?

I realize there are groups researching into artificial intelligence, not positive if any practical progress has been made. But in any case suppose a sentient computer is constructed at some point. The idea is great and all in theory but is there any practical usage for AI? Besides habitual laziness of course.

It would seem to me that there is no use besides novelty. All critical and non critical operations on the planet use computers with no more intelligence than a home pc. Even the most complex programs such as those in medical purposes have no showing of intelligence.

Are we just wasting time developing technology that has no purpose?

It was thought, in the late 1940s, that the UK might eventually need four or five computers in the entire country. There was no practical use for more than that. I place your thoughts on this in the same category.

Off the top of my head here are some of the things you could do with AI:

1) Mundane clerical work.
2) Rapid sifting of research literature assembling meaningful synthesis of topic areas.
3) Cross disciplinary sifting of research literature to identify unexpected connections between events.
4) Assessment of environmental monitoring data.
5) Companionship for solitary persons.
6) Quality simultaneous translation.

I am sure there are thousands more.

Another reason for research in AI is that, by aiming to imitate or replicate human abbilities, we may learn more about how our brain does the same job. The intirncacies of shape recognition, fai, led to the research and discovery that our brain's visual processing center haves specialyzed cells for shape-recognition, which tell between horizontal, vertical and tilted lines.

And think how great gaming would be with AI NPCs. :-D

Anything tedius or time consuming: astronomy, medical observation, government spying--- I thought we already live in an age where AI is indespensible, but I don't think of AI as necessarily Robby the Robot.

Originally Posted by Helios

But in any case suppose a sentient computer is constructed at some point.

AI is just a lofty name for sophisticated search techniques. It has nothing to do with "sentient" computers or other magical things.

Originally Posted by Ophiolite

1) Mundane clerical work.
2) Rapid sifting of research literature assembling meaningful synthesis of topic areas.
3) Cross disciplinary sifting of research literature to identify unexpected connections between events.
4) Assessment of environmental monitoring data.
5) Companionship for solitary persons.
6) Quality simultaneous translation.

1) Maybe helpful.
2) eh... what is the point.
3) '' '' '' '' ''
4) Eeeww, I hope I'm not there when this happens.
5) Aaah no. This is not good whatsoever.
6) Not sure what you mean?? :?

I have done a lot of research into AI in the last year. Its very hot stuff and is being used in a lot of high level stuff... the idea of HAL the sentient computer is long out the window, what people are working with now could be described as "life inspired algorithms"...things that learn over time from experience, adapt to different environments, etc. Its probably one of the funnest things in computer science, in my humble opinion!

Originally Posted by DaBOB

Originally Posted by Ophiolite

1) Mundane clerical work.
2) Rapid sifting of research literature assembling meaningful synthesis of topic areas.
3) Cross disciplinary sifting of research literature to identify unexpected connections between events.
4) Assessment of environmental monitoring data.
5) Companionship for solitary persons.
6) Quality simultaneous translation.

1) Maybe helpful.
2) eh... what is the point.
3) '' '' '' '' ''
4) Eeeww, I hope I'm not there when this happens.
5) Aaah no. This is not good whatsoever.
6) Not sure what you mean?? :?

DaBOB, I did not notice this reply when you posted it.

2) Rapid sifting of research literature assembling meaningful synthesis of topic areas.eh... what is the point.
I thought the point was clear. The volume of published material inthe sciences has risen exponentially. It is barely possible for researchers to stay current with work in their very narrow field of interest, and impossible to explore fully related and potentially useful fields. AI could be used to sift all publications and produce summaries of the work from a perspective relating to the individual researcher.
This would do far more than simply save the researcher time. It would provide her with insights and information that would simply be unavailable without the AI processing.


3) Cross disciplinary sifting of research literature to identify unexpected connections between events.eh... what is the point.
I am astounded that you cannot see the benefit of this. There is the old joke about increasing specialisation: we learn more and more about less and less, till we know absolutely everything about nothing. AI can break this log jam by looking for potential cross disciplinary relationships.

4) Assessment of environmental monitoring data.4) Eeeww, I hope I'm not there when this happens.
What a bizarre thing to say! Why would you see this as undesirable?

5) Companionship for solitary persons.5) Aaah no. This is not good whatsoever.
Wrong. The Japanese are already using a simple form of this with excellent results.

6) Quality simultaneous translation.You speak to a Malaysian in Lativan. His AI translates it, as you speak, into Bahasa Melayu. He replies and your AI translates it into Latvian.

My computer science prof used to bring it to the point: "There is no artificial intelligence, only natural stupidity. :wink:

If you want AI, forget the PC, it's no good. Parallel processing, of the true kind is needed. Our current serial processing PC's can never grow to AI.