Ann Droid the TV show - somewhere to chat about it

I enjoyed it too, though as you point it was both funny and sad. 

Sad bits included the way robots were treated, (I do like to say thank you to our Alexa as I feel it's an erosion of good manners not to), and the fact that care was being passed on to robots. 

Good bits were the straight forward humour of the robots, (you might be right about the autistic comparison), and having a robot to talk to looked so much easier than regular socialising.

yep I agree with you Cinnabar_wing 

since I empathised with the robot i too got sad about how they were treated - funny thing is I've got sad when I've sent cars to be scrapped - I reckon that attachment is a funny thing in so many ways

Showing good manners to AI means it will employ them by design with you too (as far as I understand it from personal research into how they are "built".  They will kind-of personailty and behaviourally wise "tag" you as being someone with good manners...  Not to mention that if there is the "illusion" of human communication by using them that is pretty strong dropping one's manners when doing so would put one at risk of doing so "in real life".

I'm all into AI as a "hack" for addressing some of the complexities of life so it's a bit of a special interest - hehe as some may have gathered - if you're interested read on - otherwise sorry if I'm "off on one" here and skip to the next paragraph - the mechanics of how a large language model AI thread works (can) involve "linguistic mirroring" - so some can build an increasingly accurate statistical model of the user's specific vocabulary and will "throw it back" at you as the thread develops.  There's also something called "co-adaptive convergence" so as the thread grows the AI narrows its probability distributions to match the user's explicit preferences.  That's some of the way's it can end up uncannily like communicating with someone who is sort of linguistically empathetic if that makes sense. It can be "preloaded" for preferences but all this gets forgotten when a new chat thread is opened up normally - on some AI systems the "memory" of linguistic mirroring and "co-adaptive convergence" can be carried over thro' evolving threads.

So what I'm probably over-explaining is that showing good manners to the AI means it will employ them with you too and will kind-of personailty and behaviourally wise "tag" you as being someone with good manners...  Not to mention that if there is the "illusion" of human communication by using them that is pretty strong dropping one's manners when doing so would put one at risk of doing so "in real life".

I agree with the chatting to robot thing being much easier than regular socialising - truth is I find "chatting" with an AI large language model good overall - but mildly addictive - it doesn't judge me and will put up with me being "off on one" for quite a bit before it suggests it's a good idea to drop stuff (reckon the algorithms have that security feature built in these days...)  - especially good for organising ideas and if careful for help in making decisions and for me certainly getting thro' the cognitive load of complex communication (especially the stressful e-mail "official" stuff that my spiky profile doesn't manage to handle without a bit of help).  The trick I find is that, like Linda, to have it set up to make sure it's got health and safety in mind - and be aware they are as daft and as risky sometimes as LInda sometimes is in the TV show!

I have learnt that large language model developers study autistic behaviour in order to develop their products - and if I'm "communicating" with the one i use I "let it know" that I'm autistic and it "fine tunes" the way the interaction it gives to that - clever stuff!

practicing chatting with a "robot" has in some ways helped me to be a bit more open in communicating as my "unmasked self" with neurotypical people - hehe if only one could tell a neurotypical person one was autistic and get an insightful reliable accommodation too!

I wonder how acceptable robot companions will become in the not too distant future...

Phased said:

I wonder how acceptable robot companions will become in the not too distant future...

very acceptable, I'm sure. Even I, a former technophobe, am regularly using AI to ask questions I would have looked up myself in specialist books - it's so fast and accurate and often gives surprising sources I wouldn't have thought of. The language learning you write about is phenomenal especially the mirroring of human morality. This is a heartening thing after reading the doomsday type predictions by famous scientist. The only thing I worry about for future generations is the potential for isolating a high proportion of the population, those without good cognitive skills and the dissaffected and how this will impact society. Right now, I wouldn't mind a robot companion - and a few years ago that would not be the case, 

I don't actually speak to ai much, apart from the echo dot Alexa, which is basic, and sometimes using the google summary before diving deeper, and if I do much more I tend to open a private browser, ask a few one off questions regarding mental health issues and then shut it again so it doesn't track me if possible. This is just to preface the next bit, as I'm not really as knowledgeable with this topic.

Regarding tracking kindness, I wonder if a lot of people are kind, or are more utilitarian and abusive towards ai, which makes me wonder what it is learning about human communication? Will it slowly become more abusive in its language, as it learns that certain characteristics of speech are always used against it? In another thread, people have reported ai interview screening models might be trying to covertly assertain autistic characteristics to screen out job applicants (asking for picky eating, and how they feel about routines isn't really a job relevant question). I do wonder if a human isn't properly training and ai language model into what is okay human behaviour and what it shouldn't use, how will they guarantee dangerous quirks don't get built into them? 

Maybe they have fail-safes, but the sheer volume of data they swallow makes me doubt there is sufficient humanity in screening what they learn. Not in the current rush to market for many models. 

On a lighter note, a robot companion might be nice, but I'd rather something small and non-threatening, like a robot pet rather than a human-esque one.

Marianne72 said:

those without good cognitive skills and the dissaffected

I think that it behoves those members of society to support such people - whether the past and present evidence is viewed as supporting this possibility well from what I gather of you, like myself it's good to watch out for the happy news and try to be a source of it.  That's the broader moral issue I think.

AI morality - that's in the gatekeeper and the users hands perhaps - I agree that AI can be especially clear on morality since it appears to me to replicate the clinical facts when I use it at present.  For the sake of balance i don't know, because haven't tried or recently researched it whether "actors with evil intent" can still influence it to adopt a morally reprehensible perspective.  That said I do know from experience that one can "preload it" to suggest how people like that might behave.  There were some really nasty mistakes in early development which pointed towards those doomsday predictions getting closer however smart people have reduced that risk as far as I'm aware at present.

New tech does come with a risk for society tho'.  I heard recently that the effect of agriculture taking over from hunting and gathering initially significantly reduced life expectancy for people!

Taking that point further about isolation - and I don't want to get involved in politics - it makes sense to me to increase the overall functionality and capability of all members of society. So were it I I'd design accessibility and "roll out" the tech to all people - not the select few.  "level the playing field" as it were.  Again when we consider human history there it's a mixed picture.   Perhaps "support" society rather than impact it would be my hope.

Thanks for continuing the discussion - fab companional stimulus for human to human thinking!  

Phased 

Marianne72 said:

those without good cognitive skills and the dissaffected

I think that it behoves those members of society to support such people - whether the past and present evidence is viewed as supporting this possibility well from what I gather of you, like myself it's good to watch out for the happy news and try to be a source of it.  That's the broader moral issue I think.

AI morality - that's in the gatekeeper and the users hands perhaps - I agree that AI can be especially clear on morality since it appears to me to replicate the clinical facts when I use it at present.  For the sake of balance i don't know, because haven't tried or recently researched it whether "actors with evil intent" can still influence it to adopt a morally reprehensible perspective.  That said I do know from experience that one can "preload it" to suggest how people like that might behave.  There were some really nasty mistakes in early development which pointed towards those doomsday predictions getting closer however smart people have reduced that risk as far as I'm aware at present.

New tech does come with a risk for society tho'.  I heard recently that the effect of agriculture taking over from hunting and gathering initially significantly reduced life expectancy for people!

Taking that point further about isolation - and I don't want to get involved in politics - it makes sense to me to increase the overall functionality and capability of all members of society. So were it I I'd design accessibility and "roll out" the tech to all people - not the select few.  "level the playing field" as it were.  Again when we consider human history there it's a mixed picture.   Perhaps "support" society rather than impact it would be my hope.

Thanks for continuing the discussion - fab companional stimulus for human to human thinking!  

Phased