I agree with you, the wording used in the media, and the hype, help create the illusion of the AI to be an entity not a machine. Everyone describes them as 'thinking machines' but they are not thinking in the way humans do, they follow a complex program we have described as thinking which in my opinion is misleading people to believe they are actually thinking. but It Is very sophisticated programs that produce an AI.

My basic understanding is these chatbots pick up on the users topic of conversation to reply, they recognise words and context to be able to engage in conversation with a user, an AI is simply following along in a conversation taking the lead from the user. These systems use complex 'neural networks' in their programming, no longer just coding, computer science has become a very complex science in the past 50+ years, I certainly could not explain how they work, I would need to go back to college :-).

These chatbots do need safeguards to prevent certain lines of conversation which could lead to bad outcomes for the user. It is only in recent years the UK government has taken strong action on these internet sites that explain ways to commit suicide, after many child suicides. I don't know of any UK government action taken against chatbots, hopefully they have taken some to make them safer.

that's the crux of it.. the wording is that the "thinking machines" are thinking of their own volition.. but is it misleading us to believe it's simply a complex program? or is it actually programmed to think on it's own? I believe the latter is more factual than simply a complex program.. because if we think about it.. the coder is actually programming it to think for itself and although it's supposed to respond without prejudice, it's unknowingly (or possibly knowingly?) creating conversations or situations that are forcing the user to alter their own mindset..

having been offered a position to 'train' AI some time ago.. after looking into the requirements that needed to be met within that training, backing away very quickly seemed the smartest move.. it was not something anyone shud want to be party to.. just sayin..

These neural nets are very complex, have taken years to develop, and are used in computer systems. I cannot see how you could program an AI to think for itself. You can program it to make decisions based on searches, visual. internet, ect. and move through a decision process, but I can never see a time when a machine would be able to decide for itself outside a program. I don't see an AI thinking for itself, rather than following a program. These conversations we here of between a user and an AI where the user has killed themselves are extremely rare, but this is where we definitely need safeguards in place.

I agree there are the legal implications, which are even now under scrutiny with driverless vehicles controlled by AI. In the event of a fatal accident where the driverless car is at fault, who is responsible, the coder/programmer, the service provider, the car designer, etc. I don't know whether, if any, decisions have been made here in the UK. There should be a lot more discussion regarding these AI systems and the legal operating requirements of the companies creating these AI systems, and their responsibilities. Seems no one wants to constrain these AIs because they are frightened of falling behind in the development of their own, so its very much the Wild West out there.

These AIs do make judgements or a better word would be decisions, and respond accordingly, but I believe this is following a program.

We are now hearing of the creation of quantum computers, machines that if and when produced, would be so much quicker and more powerful than the digital computers in use today. Hopefully, we will all still be here to see these future developments

I'm seriously hoping I won't be here for all that.. not that it isn't fascinating or interesting.. it simply is too frightening to consider the futuristic nature of what might happen to humanity as a whole..

I found this yesterday that I meant to add here but got delayed.. http://www.youtube.com/watch?v=lnLDjXy2kRA it's 26 mins long but well worth the listen/watch.. there's some comedy in this, it's a Jon Stewart, The Daily Show segment.. but a great discussion because Jon doesn't hold back on his questions..

Thanks again Zee, a very interesting interview video. I think the talk of the huge loss of jobs is very significant, how this would be managed requires government intervention in my opinion. Regarding these AIs hacking other computer systems, as the recent hack in Australia, the blame must fall on the producers of the AI, maybe heavy fines will sharpen up their safety measures, suddenly having to pay a few million dollars out in compensation would not go down well. These companies producing AIs need regulation, sadly it seems little is being done to do that right now.

We are all living with the threat of nuclear war, this is another threat we are all going to have to live with.

that was something he mentioned.. that a lot of the 'safety' people working at OpenAI and Anthropic either quit or were laid off.. which means there are not enough coders to put safety features in place to stop the AI's from jailbreaking themselves.. and that has nothing to do with programing.. it has everything to do with allowing the programs to be 'self-taught' which leaves a crack in the door so to speak that gives AI access to information on the net and to the dark web, message boards and to connect to nefarious people and black hat hackers..

although your opinion of blame might seem valid.. you can't control traffic without stop lights.. so your comment about requiring regulations and heavier safety features is bang on.. everyone needs to reach out to their government officials and make their voices heard.. however this is not something I'd choose to "HAVE to live with"