AIBU to say AI will completely change how we parent and live?

AI is a tool. Books are a tool. The internet is a tool. You use it as a virtual planner (or whatever). I’m not sure how changing the tool is changing how you live. It’s just a different method of doing the same thing.

You can ask it to double check or cite sources for it's comments, which means it will check more thoroughly and correct itself if necessary. It's what I do when I'm using it for anything technical.

• write a letter apologising to Mrs head teacher for leading a mexican wave throughout her maths lesson
• Make a character for book week using green paper, coat hangers and a bin
• Is there a law about kids walking to school UK
• Find me all private speech therapists within 25 miles check they are hcpc registered
• Which schools have the best a level results Southend *Coolest trainers for 12 year old boy

Variable but all useful up to a point

That's a perfect illustration of the shortcomings of AI. People have got into all sorts of messes because they blindly follow sat navs without questioning the reliability of the information it gives them and because they can't read a map, they don't know it's telling them nonsense.

I read an article recently about an ai experiment. It had access to an “engineers” email. This engineer had suggested replacing the ai for another. The ai knew the engineer was having an affair through his emails. The ai decided to blackmail the engineer to be kept.

Scary!

I was at a talk about ai and the speaker said the rate of change caused by the internet in 30 years is going to happen in a fraction of time with ai.

I thought we are working on nuclear fusion and making strides and that will power all this stuff in a carbon neutral way,

The personalised tutoring and therapy angle, AI as a friend/teacher, is interesting. DD in Year 11 is using ChatGPT to help with GCSEs. We have solar panels. Still selling back to the grid at peanuts.

So the question of will AI be accurate is an interesting one.

AI is a really big field and as others have said there are surgical robots, manufacturing has had robot arms for production line etc for years but most people don’t really think of these as “intelligent”.

so let’s focus on the LLM (the large language models) like chatGPT.

when we are asking are they accurate, we’re basically asking are they telling the truth. When they cite academic papers or previous legal cases can we be sure those academic papers exist?

the basic answer is no.

The only way for LLM to be more accurate is for them to have a way to compare their output to reality (*I am aware this is more complicated then it sounds). So they’d need to have a feedback loop that takes what they are saying, compares it to the world and works out if it is true or not.

this feedback loop could be fairly easily built for previous court cases or academic papers where it is a straightforward case of “check a few databases to see if they exist” but the general problem of “please check whether this statement is aligned with reality” is much much harder.

there has been a pivot recently to trying to use the A.I. hallucinations as a creative tool, as most in the field now think they cannot be eliminated.

see eg demis Hassabis in this article.
https://www.linkedin.com/pulse/demis-hassabis-strange-futures-lie-ahead-mikael-alemu-gorsky-noapf?utm_source=share&utm_medium=member_ios&utm_campaign=share_via

You sound like you think it's a choice. My work already integrate it into email, use it for summaries of meetings etc. I cannot personally opt out of that. I might not spend hours using it to create pictures of me with silly hats but I can't say that I can avoid using it for most tasks at work now.

Yes, because it will accelerate global warming at an unprecedented speed. There will be massive repercussions from you and everyone else using it to simplify your life. Crack on if you think that the pay off is worth it.

That's the concern with quantum computing. It is real and it is coming soon. It would take a modern supercomputer billions upon billions of years to crack the current encryption algorithm we use. It would take a quantum computer a fraction of a second, and they've already been developed. It is concerning.

In the series silicon valley.....spoilers...

They integrate AI into their compression algorithm and it ends up being so powerful that it cracks every encryption code ever made. So no data would be safe, no privacy, no bank accounts etc. So they end up realising they have to destroy their own creation.

@Frequency oh yes I want one if those too

Actually there are multiple other operating systems for smartphones.

https://en.m.wikipedia.org/wiki/Linux_for_mobile_devices

i will concede no-one uses them except serious unix geeks

@Octavia64 I know it's not accurate now, you can see it on some of the google suggestions but posters are talking about the future. Or will it never be accurate?

Children’s world and job experience being massively different from school/parents is not new though. when my husband was at secondary school, the school had one computer that was wheeled about on a trolley into lessons. his 2nd job after university was managing the company network for its Europe, ME and Africa divisions. If you had mentioned a computer virus to his O level teachers they would not have had a clue and thought you were being beyond fanciful.

I'm waiting for an AI/android gardener. Not the robot lawnmowers we have atm, although I also want one of those desperately if anyone has any recommendations, but I want one that will trim the shrubs, do the weeding, keep plants alive, and know when it needs doing without any interaction from me. Once they are available, I will rob a bank to be able to afford one

Eventually, AI will be a useful support in my job but can't replace what I do so no concerns there. I wouldn't use it for many of the things listed here. It has a long way to go before being considered even remotely reliable though

Chat Gpt is just a tool, like a roadmap, so you think people should simply read a road atlas while driving instead of using Android Auto?

That's what I'm referring to. The tech will become so ubiquitous that not using it will become a burden.

We’ve already had the first court case where a case was thrown out due to AI making up previous cases.

https://gunnercooke.com/ai-fake-cases-and-the-courts-a-cautionary-tale-for-the-legal-profession/

the current problem with AI is that it hallucinates (makes things up). So if lawyers use it to put together legal arguments it makes up old court cases.

a good way to lose your case.

Is AI perfect though? Just thinking of that plane that was landed on the Hudson didn't the computer suggest going to the airport?

I would be very happy with a robot chef/housekeeper

You are underestimating how ubiquitous A.I will become.

7 year old children won't need a phone or any physical device to get access to knowledge. The applications will be embedded within most devices, in a passive sense.

There won't be a carefully sequenced introduction to knowledge, because knowledge will have no value. Knowledge will become redundant, why learn to write when your voice (or literal thoughts, yes, this has really been done, a Pink Floyd song has been sequenced from just monitoring thoughts) will do it for you.

We don't have words to understand the changes coming. Just like the Right Brothers didn't understand aerodynamics, even thought they invented the airplane.

Surgeons or we won't need so many?

I can see it working in things to do with law, so rather than a lawyer reading through a ton of paper to find a precedent or caveat. It could save a lot of time, it could probably help on how to contrast a closing argument/defence etc based on what has been successful prior.

Marketing eg targeting the customer more accurately and effectively.

Will AI help us learn how to do all our own DIY?

Yes. Isn’t that rather the point of it?