Nick Abbot on LBC talking about Artificial Intelligence

[6EQUJ5]

Nick Abbot comments on what is happening in the news. This is last night's programme. After a few minutes he talks about the near future of Artificial Intelligence, it being bad news for a lot of people. To begin playing the programme click on the blue triangle in the centre at the bottom of the screen. The news and the adverts have been deleted.

https://www.globalplayer.com/podcasts/episodes/7DrfLHJ/

[Psaw]

Artificial intelligence cant do owt that it hasnt been programmed to do. It hasnt a soul, it cant choose, think, or feel. Which brings me to you 6eq, you function within the parameters of AI.

There doesnt seem to be owt else.

[griff]

Whereas with you, Frank, we just leave out the word ‘intelligence’.

[6EQUJ5]

Artificial intelligence cant do owt that it hasnt been programmed to do. It hasnt a soul, it cant choose, think, or feel. Which brings me to you 6eq, you function within the parameters of AI.

There doesnt seem to be owt else.

Your ever poorer insults are of no consequence to me but they do say something about you. You have a need to try to bring me down to below your level. But that is a substantial distance to fall and you are not capable of performing the task of initiating that fall.

My job was as a computer programmer / analyst on computers that filled large rooms.
It is obvious that computers are not sentient. But the fundamentals of Artificial Intelligence are that it can learn. Give it access to the internet and it can learn everything from everywhere. There are recent examples of two computers communicating with each other. They made their own language and after a while the connection had to be switched off because the common language created by the two computers could not be understood by anyone except the two computers. There is also mentioned on this radio programme that an AI application was given access to the internet and within a couple of hours became a Nazi. There is also an example of two computers actually communicating with each other in English and getting feisty with each other.

AI programmers write the software such that it can be free to learn and improve from its initial state. That makes it dangerous. AI software can itself write software and design computers that are better than it. It is extremely old-fashioned to believe that these AI programs are merely sets of instructions that carry out tasks that their programmers have set. They are a long way beyond that. They will soon reach the obvious stage of being unable to be stopped by being switched off because they will be able to copy themselves onto other computers worldwide via the internet. Psaw, you have a quaint old-fashioned view of what artificial Intelligence is. It is a long way beyond your simplistic view. There are programmers who have made it so. Furthermore, it is likely that AI cannot be prevented from making further enhancements because even if agreements are made to stop developments it is highly likely that work will go on in the likes of Russia, China, North Korea, USA and countless other rogue countries. AI will not be stopped and it is dangerous. That is because it already surpasses human intelligence. It can only increase this intelligence gap. Current AI will create better AI and that will create better AI etc. It is crass to believe that pulling the electricity plug out of the socket will for ever stop it. It will get around that problem by replicating itself over the internet. Nobody is going to switch off the entire internet. Even if that happened that will not stop it. The very nub of the issue will be that its intelligence will be thousands of times better than the best of us and it will survive. That is the future.

Maybe the answer to the Fermi Paradox is that a civilisation reaches the stage of being able to create AI and then does not last much longer.

[Blackdogblue]

Have seen I Robot & the Terminator… did not go well

[Chauncey]

Your ever poorer insults are of no consequence to me but they do say something about you. You have a need to try to bring me down to below your level. But that is a substantial distance to fall

I genuinely laughed out when I read that 😎

Computers getting ever more advanced and speaking to each other surely is a genuine worry for future generations,it's possible that we've created the monster that will finish off mankind,if computers get to the point where they realise that they no longer need human intervention and infact we are holding them back then we are in trouble,it's a very interesting what you say about talking with other and if they have the ability to develop their own language then the first steps forward have been taken by them all be it unintentionally

What did you mean by the Ai appliance had become a nazi 6EQUJ5? It's thought provoking stuff

[6EQUJ5]

I am not sure about the minutiae of the AI system becoming a Nazi. It was connected to the internet to gather all information about everything. The information is from Nick Abbot's program on Saturday night as per the link. He reckoned that the AI system gave the opinion that Hitler should have been more successful.

Churchill thought the same as the AI system. During WW2 the UK intelligence services came up with several schemes to kill Hitler. One was quite amusing. Hitler was charismatic but not sane. He was known to occasionally chew rugs. The UK intelligence services planned a scheme to get a poisoned rug delivered to Hitler. However, Churchill absolutely forbade any scheme to kill him. The reason was that Churchill regarded Hitler as a poor war leader. If Hitler was killed then the next leader may have been much better at the job. All the German high ranking people were Nazis with the same aims as Hitler.

[Psaw]

Yes they are. Switching off the internet sounds fine, computers are made of metal, its biggest failing, a strong magnet would foul it up. No electric no existence.

While the electricity is down, go for the brain, the heart. Switch on the electric with something nasty in the electricity. Unless its be programmed to be cunning evil, it wont be. If AI learns, then it will learn it cant win if it challenges humans.

If what you say is true, im sure AI could cause mischief, we already know, fatal for AI.

If it tries to knock out human kind it had better do it in one go, a ten year old child will be able to defeat it, never mind a fireman with an RPg, an F16, or Jeremy Corbyn.

Human intelligence is off the scale, you wouldnt think it though. Youve insulted me by the way. You started it, saqwatch. AI will only ever function through that which constitutes a body, which will be mettalic.

Throwing up after a night on the town will fuse it.

If humans are connected to electric stuff watch out.

[Chauncey]

I am not sure about the minutiae of the AI system becoming a Nazi. It was connected to the internet to gather all information about everything. The information is from Nick Abbot's program on Saturday night as per the link. He reckoned that the AI system gave the opinion that Hitler should have been more successful.

Sorry,I've got to admit that I didn't listen to the link

If the Ai system had been told to gather information about everything then surely it would have realised that Nazism was wrong and wouldn't work mainly because there was more good countries than bad,plus the economic power of Germany and Japan is now massively superior to the countries that it fought against,if I was the system and learnt about everything when it would be that war is ridiculous and start developing 50cc Hondas that would take over the world?

Just playing devil's advocate btw

[6EQUJ5]

Sorry,I've got to admit that I didn't listen to the link

If the Ai system had been told to gather information about everything then surely it would have realised that Nazism was wrong and wouldn't work mainly because there was more good countries than bad,plus the economic power of Germany and Japan is now massively superior to the countries that it fought against,if I was the system and learnt about everything when it would be that war is ridiculous and start developing 50cc Hondas that would take over the world?

Just playing devil's advocate btw

I very much doubt that the AI system has any concept about right and wrong. It is connected to the internet, gathers information and may make judgements about success, failure regardless of morality. At the moment it merely answers questions and responds to requests to write stuff based upon what it knows from being continuously updated with what is on the internet. To the AI system being a Nazi would merely be a cold clinical judgement. I do not know the details of it becoming a Nazi, that may be an exaggeration. With the entire internet for the AI to consider is there more truth on it than lies? Is there more good on it than bad?