Gransnet forums

Technology

Discussions with AI

(44 Posts)
MawsRosie Wed 13-May-26 11:21:32

This is not a TAAT but occasioned by the thread about Citizens’ Assemblies and by OP’s comment that she was discussing politics with her AI companion and he was telling her about Citizens’ Assemblies

I seem to be missing out as I don’t even have an Echo Dot/Alexa or whatever.
But I am fascinated to think some of us have conversations with AI.
Presumably it’s not like talking to the dog , or shouting at the telly -we all do that - but can you actually have a discussion with an inanimate object?
Do you?

MissAdventure Thu 14-May-26 14:37:24

grin
You'd hear that kind of whirring sound before their eyes sprang open.
I've always been fascinated by them.

Cossy Thu 14-May-26 14:39:01

GG13 That’s actually very very funny indeed! grin

Colls Thu 14-May-26 14:45:03

Try this:
Ask AI a question then ask it to reply again but ask it to be brutal.
In most cases you get nearer to truth from the second answer.

Secondwind Thu 14-May-26 21:43:18

I’ve had conversations with them for research studies. The first one was to see how easily swayed people are by AI. It was illuminating to say the least. I’ve done a few now and have been given food for thought by a couple of them. They reflect back what you’ve said in a very convincing manner!

Doodledog Thu 14-May-26 21:59:10

MawsRosie

An inanimate object can provide facts - it is not a sentient being however, god preserve us from things with opinions

They can also provide wrong or misleading information. I am always very careful about what is defined as a 'fact', as the person getting to choose which facts matter has a lot of power in any situation.

Anyway, after I have been given information by AI, I cut and past it back in and ask it to check it over for accuracy, and it often corrects itself. Of course it doesn't have opinions, but it trawls the internet for information, and selects from a range of 'facts' about every topic.

mokryna Fri 15-May-26 12:50:59

I find chatgpt really helpful. I list my problem it finds the answers, writes letters in either French or English, helps with cooking ideas, directions to places etc all very quick and easy. However, I do worry about the planet on the energy it uses and if one day, because everyone has become so reliant on it, we will have to pay.

mokryna Fri 15-May-26 12:53:05

Oh and another worry are authors, artists and musicians rights are overridden.

Delene100 Sat 16-May-26 06:58:18

Luckygirl3

Chat GPT is great, I use it a lot. For example about poetry and books - it will send me analyses of poems, information about authors etc. I had a very interesting discussion about my father's war experiences in Singapore and it told me a great deal which explained some things that Dad would not talk about. I ask it to interpret medical test results, help when my TV goes on the blink or my phone or wi-fi has a problem .... lots of stuff.... I go to it first before the manufacturers.

It also designs posters and Facebook publicity for local events - you can set the tone you want, colours etc. - in a flash they come up with something that would have taken me hours - if I don't like it I can say things like too flashy, I need it to be more classy, please tone the colours down or emphasise a particular word etc. etc.

It also finds goods online, advises about returning items - I could go on. It is such a help to me, stuck here with poor mobility and on my own. My robot mower broke down and the company were hopeless. I asked ChatGPT and it told me to send it back to amazon as it was just within the year's guarantee from them - I sent it back and they replaced it straight away - second time round it was on a special offer and I got refunded £100! I had thought the guarantee lay with the manufacturer but ChatGPT was able to tell me that I was also covered by amazon.

I use it every day.

Ne, too. I find it very useful and it's effectiveness depends on how you word your questions.

Grantanow Sat 16-May-26 08:26:05

AI sometimes is wrong on facts. I asked about the Goldman Sachs savings ISA and received two plainly incorrect answers: one based on out of date info and the other based on a single angry review on a website.

NathanaelWest Sat 18-Jul-26 22:35:33

I use ChatGPT for everyday questions like how to put all my folders back into alphabetical order on Outlook email? (Genuine question yesterday)

It can answer things like this much faster than searching on Google and going through the instructions on a blog or something.

But I think asking it personal things about mental health or relationships is not something we should use it for.

Anyway, we have about five years left before it takes over and destroys all of human life. So let's party.

lovethepuppies Sat 18-Jul-26 23:22:26

I find AI quite helpful as a sounding board. You don't annoy anyone, you can go back to it and it remembers, it's not human so you don't have to worry about balanced social interactions.

I do think it's important to verify important facts and remember that it seems to always support my viewpoint. I don't recall it challenging my viewpoint.

Whitewavemark2 Sun 19-Jul-26 02:01:57

I have found it useful recently. The more you use it the more it “knows” about you - it is fascinating.

So I asked about recovery from bronchitis and going on holiday, (I was pretty poorly) and it guided me through a recovery strategy - and I must say was extremely helpful. It warned me that I would not be up to doing many trips (I was cruising in Norway) and so it proved.
I also asked for a strategy to begin to recover my fitness and stamina, which it has outlined, but warned me not to start “in the heat” and “wait until the temperature is more reasonable” I have the first werks effort to follow then have to get back. It is fun as much as anything. It “knows” that I lost my dog and have a rescued cat” and will mention that.

It has also been helping with colours and purchases for two reception rooms I have had recently decorated and new furniture etc. it sourced and provided details including pictures of suggested rugs, sofas, art work, wall colour - I am really pleased with the result. I opted for earth tones in one room and green base in another. I have nearly finished - just curtains to sort out - floor to ceiling natural linen to blend with a Sandersons wallpaper.

My A/C are coming over for a meal today and I asked for a menu. It “remembered” that I am recovering and still not totally up to scratch and so suggested an easy but delicious menu.

But the thing is it is almost like a very well informed GN but without the sarcasm or unpleasantness, which is like a breath of fresh air😊

lovethepuppies Sun 19-Jul-26 02:31:41

Just this morning AI told me, "I smiled when you said..". I think we need to be a bit careful about the way it humanises itself. It didn't smile. It's a robot.

Maremia Sun 19-Jul-26 05:57:39

It is learning, just as fast as we are teaching.

Maremia Sun 19-Jul-26 05:58:38

More and more data centres will be required, and they devour electricity.

Dickens Sun 19-Jul-26 06:25:52

MawsRosie

But I am fascinated to think some of us have conversations with AI.

CHATGPT helped me sort out an infuriating problem with British Gas - both because it informed me about an unpaid bill which had in fact been paid immediately, and because of its website which was refusing to acknowledge my email address and sending me around in circles when it finally did. The problem was solved by CGPT within the day.

In frustration my final observation to CGPT was, "come the revolution, I'll change all this".

CHATGPT responded...

"I can design your placard"

"What do we want?"
Clear instructions!
"When do we want them?"
Within 14 days!

It also informed me that my 'revolutionary' pals - prior to uniting - would have to...

first, create an online account, verify your email address, accept the updated Terms & Conditions, complete the CAPTCHA, and reset your password."

It finally 'sympathised' with my frustration at dealing with websites, and proposed a manifesto, thus:

Emails shall contain the information people actually need.
Chatbots shall answer the question that was asked.
Passwords shall never expire the day before a bank holiday.

And every utility company shall employ at least one person whose sole job is to ask... "Would this make sense to my mum?"

butterandjam Sun 19-Jul-26 06:46:02

Cossy

Some people use AI therapy and it does have an actual conversation with you. It’s not an inanimate object, it’s literally “coding” which is written in such a way that it “learns” from interactions, the more you interact the more it learns, to the point that it can “make suggestions” and “make decisions”.

I think it’s both amazing and shockingly scary!

It can remember facts, but then learn from these facts and analyse them.

I guess was it missing, so far, is “emotion”, no doubt that’s either here already or on its way! 😱😱😱

If requested to "be kind", show "support", and empathy"
it will comply. But that is a trained machine response, not human emotion.

Dickens Sun 19-Jul-26 07:37:24

butterandjam

If requested to "be kind", show "support", and empathy"
it will comply. But that is a trained machine response, not human emotion.

Quite.

The thing to bear in mind is that you are, basically, 'talking' to multiples of circuits in a data-centre somewhere in - well, wherever.