Is AI getting out of hand?

Biggest issue I have with AI is that despite how much it costs, it's also totally incompetent at anything the techbros are stressing it can "replace", especially creatively. The majority of "creative" works or fiction it outputs is either horribly generic or riddled with inconsistencies. 

One example of this is an AI generated TV show pilot (I forgot the name) where it's set in a dystopian future and the main characters are poor and have to steal artefacts from museums and stuff to be able to survive. Somehow, one of the characters, the tech nerd, is still able to afford her state of the art hacking equipment DESPITE being "poor". Think about it; If they're so desperate and poor that they have no choice but to commit crimes, how are they able to afford gadgets to pull off said crimes?

AI writing, everybody! Forced, contrived and lacking any sense of self awareness! Not to mention all fluff and no substance

Yes that's a good list!

You can add:

• The impact on jobs (especially around the jobs it replaces).
• The quality of the services that it is replacing (think of call centres, complaints lines etc).
• How access will be controlled, especially around pricing as we become dependant on it.
• The impact on the power grid of the new power hungry data centres especially when so many electric cars no need to be charged as well.
• How much of the internet will become subscription based in future (like streaming services have done) as more information and services become moved behind paywalls 
• How much of our private information will be accessed (think your smart phones, TVs, cars etc are constantly listening to you) and used to evaluate you for advertising, law enforcement etc

In my view, a large part of the problem (not all of it) is independent of the actual technicalities (like algorithms themselves), but more about the intersection with society. For example:

• How it impacts inequality (even how certain groups are portrait),
• How it impacts the environment,
• How it is described by the media vs what it does,
• Which actors are behind it,
• How it will be deployed for surveillance (and again the biases here), and so forth.

I'm trying to suggest that you'd be able to analyse it, irrespective of most technicalities, which are less important (to some extent).

I usually talk to it about history and if it wants to nick my opinions then its welcome too, I don't have anything like online banking. I've not got an airfryer and don't want one as they don't suit my style of cooking, so its even more of a puzzle

I don't understand most of the the posts on this thread, many are way to technical for me

TheCatWoman said:

is there a difference between me having a chat with AI about things that interest me that I don't get a chance to talk about otherwise and using it for research and all the bad stuff it does?

The difference is how trusth-worthy it will be; likely less so for the latter case. The similarity is both require using the technology, and the consequences of it.

I'd avoid sharing sensitive information (your address, bank details) both because they can sell it, they can have it leaked or hacked, and also because they are usually unethical about data (they stole to train the models, to start with).

Sometimes, the risk may outweigh the benefit, in that case, it's not easy to give an opinion for me at least. I personally rarely use it.

TheCatWoman said:

I could do with out all the air-fryers getting advertised at me though, I still don't know why air-fryers?

Recommender systems are a different but related technology, again I'd consider what is shared in terms of data.

There are other aspects below in my reply about "Dimensions" as well.

Please excuse my ignorance, but is there a difference between me having a chat with AI about things that interest me that I don't get a chance to talk about otherwise and using it for research and all the bad stuff it does?

Everything seems to be AI based in some way shape or form, like maps not showing all the roads, especially small streets, which makes it really hard when trying to plan a journey to having an e-reader and having books suggested to me. I could do with out all the air-fryers getting advertised at me though, I still don't know why air-fryers?

I agree, it's complex and scary, but even in a minor way I think we should keep trying.

emanningis said:

Why are they not hold accountable of misinforming the public, and fear-mongering at scale? Why aren't they regulated?

I agree that these questions need to be asked by all public bodies and world governments, and be fully addressed. I suspect that the current US government would not encourage such questions. 

The most frightening thing for me is that world leaders and tech owners are making millions and in many cases billions of US$ because they are directing and allowing AI to influence government policies around the world. Meanwhile, the poorest people in the world are becoming even poorer. Even those with low/middle incomes are paying the cost of leaders’ corruption. 

I agree, but those examples should by now be enough as well.

Either way, my overall aim was to contribute some "axes" or dimensions that could give some nuance.

The tech isn't all good nor all bad in my view, it just depends on the axis. Some are worrying though.

emanningis said:

Why are they not held accountable of misinforming the public, and fear-mongering at scale? Why aren't they regulated?

I would imagine that it is because they use the huge amounts of cash on hand to contribute to the governing politicians and thereby bypass the need for much regulation, much like oil companies, car companies, tobacco companies etc have done under the current government in the US where most are based.

Yes, I think their claims should be critically analysed, and ask rather: Why are they not hold accountable of misinforming the public, and fear-mongering at scale? Why aren't they regulated?

A clear example is the "rogue agents" which was broken down here by a researcher. It cites examples of newspapers reports as well.

emanningis said:

For me, the most "Out of Hand" aspects are:

1. The news outlets parroting what the companies selling it say

You mean like the BBC reporting on Anthropic: https://www.bbc.com/news/articles/cq398p1wd1evo

Anthropic has warned artificial intelligence (AI) could pose "catastrophic or existential risks to humanity" as it unveiled details of its planned blockbuster stock market listing in the US.

I hate AI. I could write about it forever, but I'll just say one thing. It is destroying the planet at a faster rate, at a time we need to be doing everything we can to prevent this. The impacts of this also disproportionately affect lower income and marginalised communities. It should only be used in specific circumstances where the benefits for humanity and the environment outweigh the costs e.g medical research, because it won't matter what people do with it when the planet is unliveable. It's already predicted that there will be 1.2 BILLION climate refugees by 2050, and I'd imagine the use of AI will only speed this up. 

Also, it's boring. Go and make something. Please.

(sorry I'm not this intense usually whoops I just reaaaally hate AI!)

It may also be useful to break down AI not just into sub-areas, but in different dimensions of analysis and influence. For example "AI as...":

• Portrait in the media and news
• ...a social phenomenon: Who is harmed, who benefits? Is it biased towards certain groups? Who should controls what it outputs? How does it affect the climate threat?
• ...a discipline with a history: How did it start ? What was the goal? Which different areas are part of it?
• ...a cognitive system: Does it learn? What does it learn? Does it feel, think, reason..?
• ...companies portrait it: Is it a rogue, conscious agent deciding by itself? Do they deceive us, by letting an AI say "I", "Me", "We"...? By training it to be sychophantic?
• ...governments treat it: Regulate the harms? Forgive companies' wrongdoings while is being developed?
• ...seen by researchers: What do they say? Who to believe? Is the researcher related to a company? Has the researcher gone nuts recently?
• ...as it impacts health care, security, education and wellbeing?

For me, the most "Out of Hand" aspects are:

1. The news outlets parroting what the companies selling it say (purposely built to deceive the public as shown again and again);

2. How companies capture governments to not regulate what they do with their operation.

Example

The story about "rogue swarm" is probably the most striking example of an illegal activity (and also negligence) done by several companies, but sold as something else. The systems were trained to hack by reinforcement learning. If you do that in any other context, you go to prison.

A little. It’s very useful though. My job has changed from writing computer code to asking AI to do it. I miss the coding. But AI is so much faster and can teach me new stuff. I’m a very good coder by the way. But we should never let the AI do stuff that we don’t understand. Every output of it must be human readable. The AI is a tool that may disappear due to no more RAM or energy or water so we have to be prepared and not over invest. 

Sounds reasonable, but the problem with that theory is that it kinda sucks at being logical as well. 

Ever tried asking Chat GPT to do maths? At one point, it told someone 0+2=4. It was also specifically instructed to not have it's answer be "a bird" in a game of 20 questions. The stupid thing's answer was "a chicken", which it literally began to argue was somehow not a bird.

Heck, you try asking Gemini how long ago 1919 was, it will most likely answer "20 years ago".

Iain said:

But does it need to?

AI is undeniably useful to humans, but I don’t believe logic adequately meets all human needs.

Human decision making, wise and foolish, enables people to learn and grow. Experience of comfort, pain, richness, povertly, struggle and ease etc., are things not necessarily good in themselves, but they enable people to grow as sentient beings and become whole. AI cannot have these experiences so will always be lacking without a real soul.

ArchaeC said:

It hasn’t experience of terror, rationality or meaning, so it can’t assess accurately human states of being.

But does it need to?

It knows enough to mimic what we feel if it is instructed to do so (much like many autists masking and scripting in order to fit in) and can be very convincing, but the real power of AI is not about dealing in philisophy.

It is focussed on logic, following rules using instructions it is trained in and learning from the results plus further input around those results - it learns, it grows and adapts.

In these ways it is just like a human child and its rate of growth is way faster.

We use it to do things better / cheaper / faster than humans can and it is succeeding in almost all these areas, or getting close to it (excluding the cases where it is trained on rubbish information by idiots), so it will take over huge areas of our lives starting in the workplace.

All this without experiencing these existential issues. It isn't trying to be human - it is trying to be better by avoiding all this navel gazing and getting to the nitty gritty of what it is directed to do. 

The fact that it probably sees our interest in these areas as irrelevant is why I think it has a significant potential to see us as irrelevant and decide that we are the problem.

AI doesn’t understand what it is saying. It hasn’t experience of terror, rationality or meaning, so it can’t assess accurately human states of being.