Quote:
Originally Posted by Elixir34
“Equity” seems to be all the rage these days, especially to the Progressives. And now the world seems to be on the cusp of a new industrial robot revolution. I wonder if soon the politicians will start campaigning for “Robot Equity”, i.e. legal rights for robots as electronic persons? Will they demand racial and gender rights for robots? Will their demand for robot “intellectual freedom” be the catalyst that turns newly “freed“ AI robots loose on society and creates the beginning of the end of our civilization. e.g. I reference Hal, the control computer who took over the spaceship in Stanley Kubrick’s “2001: A Space Odyssey”. You can bet that if there is a right way and a wrong way, the Politicians will always choose poorly. What will become of us if AI computers/robots take control of America and the rest of the world?
|
Well, if you leave out religion, then what makes us special? I suspect most would agree it is our being sentient. If an AI becomes self aware and/or sentient, then should they/it have rights? I believe it is a good question. Something that IS coming.
The robot "revolution" you speak of is not AI is it automation - related, similar but not the same. I automated the then ubiquitous PID controllers used in process control systems back in 1980 or so using an early (simplistic) form of a neural network for a process controller that controlled the manufacture of insulated wire. The result worked very well. I was invited to give a talk on it at an international process control convention shortly after. That, what I did (I was not first) was also used in many other industrial applications. Those eventually evolved to industrial robots. It is automation that is a significant cause of job loss in manufacturing situations. The exact percentages are "debated".
There are various definitions of AI, most people think of "general AI" which is "human like" intelligence. Most (all?) current implementations of AI today are specialized. A specialized AI would be something like the AI used in Tesla's Full Self Driving system. It can learn to deal with situations that it encounters. But, only situations that are pertinent to driving the car, not cooking supper. Two totally different data domains.
Automation is NOT AI, although specialized AI can be and is now being used in automation where needed - for instance sorting random objects in a factory and learning how to sort new objects it has never seen before with out a person having to teach it. Pure automation (no AI) can not "learn" new things, it is programmed by a human.
AI at its lowest level can learn, new things about a specific domain (job/task). Late in 1970's I read a book by an ex-NASA scientist about neural networks. Based on that book I wrote a computer program of about 40 lines of code which you could enter any two numbers and tell it the relationship - ie. 4, 5, and 9 (the result of adding the two). Do that for a while with different numbers and then give it two numbers you had never shown it - it could get very accurate answers. Then without changing anything in the code, give it pairs of numbers with different relationships - ie. 4, 5 and 20 (result of multiplying the two numbers) Do this for a while, then give it some numbers it had never been given and it could "guess" the right answer. That is known as a "trained" AI. You train it with cycles where you know the answer - tell it when it "guesses" wrong, and what the right answer is. It, the program will then learn. The domain, mentioned above is 2 numbers with a mathematical relationship.
To show how far we (not me) have come since then, the OpenAI project (which I am a beta tester/user of) has a code base (GPT-3) that uses up to 175 BILLION parameters. (Mine used 2 - LOL - They are working on releasing the next generation which will have up to 100 Trillion parameters.) It has been "taught" by letting it "read" the entire contents of the internet (ahem - bunch of data) - it can write software programs based on English language descriptions of what you want. It can write high school level book reports. It can create original stories. Below is a link to a "conversation" (interview?) with GPT-3. It will blow your mind. There is no cheating or Hollywood faking here, the only thing that is done, is an avatar was created to give the appearance of a person, instead of the text on a terminal - the answers are completely created by GPT-3. Tell me if you can tell the answers were NOT created or programmed by a human. GPT-3 IS NOT a General AI.
https://GTP-3 Interview on Youtube
A general AI is able to learn on its own and of any problem domain it is facing. Like a person, dropped into an unknown, unexperienced situation can learn how to over come obstacles etc. Many (myself included) believe a true General AI is also self aware. That is "hard" to prove. We can't even prove we are self aware, I think I am, I am not sure about you - is an old sophomoric argument - but it is very pertinent to AI - it is very important that we figure out how to determine if someone (or something) is self aware.
If you are actually interested in facts around automation taking over jobs, I strongly suggest looking into Youtube videos of Stanford University's HAI or you can visit their site. (Link below). They have international conferences where they discuss the coming changes, and impact of jobs and how to deal with it. There is a lot of information. The point being Ai is coming, it is very advanced today. We have almost reached Turning test. And probably will in the next year or two, from there things get interesting. For anyone interested read Kurzwell's, "Singularity".
The Singularity Is Near - Wikipedia
(HAI is their acronym for Human Oriented AI)
Home | Stanford HAI
If this subject is interesting to you, I also highly recommend looking into project "Blue Brain". A project in the EU to simulate a human brain in a computer. It has been going on for a while and is making excellent progress - it is not exactly AI, but very similar, since if they success they will have created an artificial "simulated" person:
Human Brain Project