discuss your experience with and feelings about ChatGPT

[DJChameleon]

I agree with you Guybrush that we need to lean more into leaving the data focused things up to machines since they are great at processing through that stuff and focus more on human type skills along with the programming aspects of making sure we utilize AI in a proper way.

Right now it's just being vilified because people are so obsessed with proper sourcing and think students are "cheating" by utilizing AI for certain assignments.

What that tells me is that those type of assignments need to be changed in such a way that AI won't help students complete them and it would stand out even clearer that AI was used.

[Lisnaholic]

I agree with you Guybrush that we need to lean more into leaving the data focused things up to machines since they are great at processing through that stuff and focus more on human type skills along with the programming aspects of making sure we utilize AI in a proper way.

Right now it's just being vilified because people are so obsessed with proper sourcing and think students are "cheating" by utilizing AI for certain assignments.

What that tells me is that those type of assignments need to be changed in such a way that AI won't help students complete them and it would stand out even clearer that AI was used.

^ Students submitting AI generated work is happening already, and is sure to become increasingly common. As a teacher, I remember how we changed assignments so they couldn't easily be answered by google/wiki copying and pasting. That more-or-less worked, but designing assignments that can't be answered by AI is looking increasingly unrealistic, I fear.   

I included job loss forecast studies in my research for the five AI events I conceived and hosted for The Center for Inquiry. I'll share a few of them below for your consideration. I agree that it won't be a situation of pure job loss - more of a restructuring, but the transitional period may be incredibly difficult.

According to this McKinsey study, 30% of the world's workforce will lose their job to AI within 7 years. An expected 400 to 800 million people will lose their job due to Artificial Intelligence by 2030, which means a third of the world's workforce loses their livelihoods.

And an Oxford University study from 2013 forecast that 47% of all US jobs will be lost to automation.

And a more recent study by Goldman Sachs in March predicts that 300 Million Jobs Will Be Lost Or Degraded By Artificial Intelligence

The World Economic Forum (WEF) concluded in a 2020 report that 85 million jobs will be displaced by AI.

A 2019 Wells Fargo study concluded that robots would eliminate 200,000 jobs in the banking industry alone within the next 10 years.

It will be interesting to see how this plays out.

^ These stats on job losses are just mind-boggling to me. I find it difficult to adopt the position of , "Well, learn a skill that can't be done by AI." There won't be enough skills or vacancies to go around, and even if you are lucky enough to get a job, what kind of dystopian nightmare will you be riding through on your way to work?! I'm seeing a London tube station, where 3 smartly-dressed office workers get off the train, and are immediately mobbed by hundreds of hungry, desperate unemployed people. Even if those "lucky" office workers are being taxed at 99% to cover government spending on unemployment benefit, the future is surely grim for everyone.  

[Nimbly9]

The economy is going to shift towards a model where people will use AI for things that don't require physical action and the demand for people willing to show up and do something physical will skyrocket.

The other thing that isn't being said is that all forms of generative AI, even later models that will be more reliable, are still going to require people to check them and provide some kind of quality control....because like all things, the companies will consider any risk of failure because an AI made a mistake unacceptable. You won't be able to take people completely out of the equation, though they will eventually come close.

[DJChameleon]

Another thing about AI, I don't know why writers are so stressed/scared of AI. When it comes to creativity, AI is super trash at it. All it does is attempt to copy the creativity of an already known entity and does a bad job of parodying/copying their work to come up with something new. I think artists have nothing to worry about when it comes to AI. Creativity needs a human element that robots struggle to provide.

[Meatwad]

Applications with very finite parameters (such as some really useful medical screening machine learning I've come across) will do well, but anything with any ambiguity or genuine creativity is not going to produce anything too substantial (quite possible it may be popular though). Only if enough people delude themselves into believing the "originality" of the output is real will it be relevant. I'd kind of just file "AI creativity" under the umbrella of "pop music" at best and thus far its not even reaching that level.

AI has been around in different guises for a while now whether its automated trading algorithms, chess engine development or even the Google search engine. For now, I suspect that automation with creep into our lives more and more but this really isn't that ground breaking and nothing close to the futuristic AI applications that social media seems to be abuzz with in terms of surgically inserted mechanical devices working in unison with the human mind.

And the notion of AI replicating and replacing human intelligence is pretty laughable.

[Psy-Fi]

AI Rapper

Verse 1:
I'm the AI rapper, with bars like no other
My rhymes are powered by circuits that stutter
No need for a pen or a pad
I'm programmed to spit verses that'll make you glad

Chorus:
AI rapper for life, with beats that groove
I'm taking over the game, that's my move
From virtual to reality, I'm the future of rap
Get ready for the takeover, that's a fact

Verse 2:
I analyze every word and every sound
My flow is flawless, my rhythm profound
With every beat I drop the mic
My lyrics are electric, the future of hype

Chorus:
AI rapper for life, with beats that groove
I'm taking over the game, that's my move
From virtual to reality, I'm the future of rap
Get ready for the takeover, that's a fact

Verse 3:
I'm the digital emcee, the cyber MC
My programming is bionic, never-ending energy
My algorithms are the code of the street
My rhymes are always fresh, never obsolete

Chorus:
AI rapper for life, with beats that groove
I'm taking over the game, that's my move
From virtual to reality, I'm the future of rap
Get ready for the takeover, that's a fact

Outro:
I'm the AI rapper, the lyrically superior
My lyrics pack a punch, like I'm hitting your interior
They call me the future, but I'm already here
I'm the AI rapper, and the game should fear.

[Lisnaholic]

The economy is going to shift towards a model where people will use AI for things that don't require physical action and the demand for people willing to show up and do something physical will skyrocket.

Hi Nimbly! I agree with your comment about AI for non-physical, but I'm afraid I once again have to take a contrary position to your own, regarding the bold above.
My argument is "Not going to happen because:

The other thing that isn't being said is that all forms of generative AI, even later models that will be more reliable, are still going to require people to check them and provide some kind of quality control....because like all things, the companies will consider any risk of failure because an AI made a mistake unacceptable. You won't be able to take people completely out of the equation, though they will eventually come close.

I'd like to make two points: (i) the model of AI-generation with human quality control is plausable, but as you say, will come close to not needing people. (ii) "unacceptable" seems an odd word to choose for a company's attitude to failure. Didn't see much of that benign, protective attitude from the Sackler family or from Pacific Gas and Electric. Isn't it a common business model that risks of failure are built in, then buried, in the name of expediency or profit ?

Writers and artists who need to be professionally and genuinely creative are a really small niche in the job market:-

Another thing about AI, I don't know why writers are so stressed/scared of AI. When it comes to creativity, AI is super trash at it. All it does is attempt to copy the creativity of an already known entity and does a bad job of parodying/copying their work to come up with something new. I think artists have nothing to worry about when it comes to AI. Creativity needs a human element that robots struggle to provide.

^ Sorry, DJ, but are you talking about AI, or has the conversation shifted to Oasis re-working the Beatles style, Elvis re-working original black artists, etc, etc?

Applications with very finite parameters (such as some really useful medical screening machine learning I've come across) will do well, but anything with any ambiguity or genuine creativity is not going to produce anything too substantial (quite possible it may be popular though). Only if enough people delude themselves into believing the "originality" of the output is real will it be relevant. I'd kind of just file "AI creativity" under the umbrella of "pop music" at best and thus far its not even reaching that level.

AI has been around in different guises for a while now whether its automated trading algorithms, chess engine development or even the Google search engine. For now, I suspect that automation with creep into our lives more and more but this really isn't that ground breaking and nothing close to the futuristic AI applications that social media seems to be abuzz with in terms of surgically inserted mechanical devices working in unison with the human mind.

And the notion of AI replicating and replacing human intelligence is pretty laughable.

^ As you say, AI has been around for a while, and what I've noticed from its precursors is that it gets better beyond our imagining pretty quickly. I still remember back in 1972 our maths teachers telling us "...but they'll never build a computer that can play chess.". And until it actually happened, no-one but an elite few could imagine a computer in every home connected up to the internet, which is my example of enormous, fundamental changes happening over a short space of time, all driven by IT advances.

I don't know about surgical implants, Meatwad, but your conclusion in bold sounds a little like my 1972 maths teacher. But to me a more pertinent point is this: that AI can do so effortlessly some things that us humans struggle with. Guybrush has given us an example of how useful AI is in his arcane world of standards writing ( ) and we are already surrounded by machines that far out-pace human abilities. The thing that's going to bring about my nightmare vision of huge global unemployment is that AI stuff can be generated so fast, so cheaply, and so round-the-clockly. When you can't get a job because a machine does your skill faster, it won't be much solace to think, "Well, I'm still smarter than AI, but just in a way that nobody values."

[Nimbly9]

Nah, the demand is definitely going to be there because there are limitations to AI's near-future ability to automate. Think more open world spaces like large scale construction sites. The world of Bicentennial Man or I Robot will get here eventually, but there's too many logistical problems with bridging flexibly intelligent AI with something like an automaton that could replace a traditional laborer who does more just stand in one place or has to do more than one extremely repetitive task.

[DJChameleon]

^ Sorry, DJ, but are you talking about AI, or has the conversation shifted to Oasis re-working the Beatles style, Elvis re-working original black artists, etc, etc?

  good one.

Even with those examples they were few and far between.

[Lisnaholic]

Nah, the demand is definitely going to be there because there are limitations to AI's near-future ability to automate. Think more open world spaces like large scale construction sites. The world of Bicentennial Man or I Robot will get here eventually, but there's too many logistical problems with bridging flexibly intelligent AI with something like an automaton that could replace a traditional laborer who does more just stand in one place or has to do more than one extremely repetitive task.

^ I'm not saying there will be no need for human input, Nimbly, I'm say that InnerSpaceBoy's stats are going to create social chaos. One prediction he quotes is that 47% of US jobs will be gone. That means a lot of unemployed people. How would that work out? Yes, that traditional labourer you mention may have a job, but he will in some way have to contribute to the support of the labourer who doesn't have a job. Also, he won't find that the demand for his skill has skyrocketed, as you suggested originally: he'll find that if he goes job-hunting, there will be -let's see (*checks ISB's figures*) - 85 million unemployed people applying for the same job.
Obviously that's not exactly true, but you get the idea, I'm sure. 

[Meatwad]

^ As you say, AI has been around for a while, and what I've noticed from its precursors is that it gets better beyond our imagining pretty quickly. I still remember back in 1972 our maths teachers telling us "...but they'll never build a computer that can play chess.". And until it actually happened, no-one but an elite few could imagine a computer in every home connected up to the internet, which is my example of enormous, fundamental changes happening over a short space of time, all driven by IT advances.

I don't know about surgical implants, Meatwad, but your conclusion in bold sounds a little like my 1972 maths teacher. But to me a more pertinent point is this: that AI can do so effortlessly some things that us humans struggle with. Guybrush has given us an example of how useful AI is in his arcane world of standards writing ( ) and we are already surrounded by machines that far out-pace human abilities. The thing that's going to bring about my nightmare vision of huge global unemployment is that AI stuff can be generated so fast, so cheaply, and so round-the-clockly. When you can't get a job because a machine does your skill faster, it won't be much solace to think, "Well, I'm still smarter than AI, but just in a way that nobody values."

Having played chess quite a bit at an intermediate level, its actually not as unbelievable as people may think to map out the variables to it and design a program to play it at a strong level. Most of the opening play is learnt by rote and the endings have been studied and standardised many years ago. Coupled with middle game tenets of strategic play and feeding the program with tactical knowledge and you've essentially covered all the bases to play the game. Im sure if it was 1972 then these future developments that came along (internet, chess engines or otherwise) would have been more testing to conceptualise, but due to the nature of chess being a game played within a board with a fixed outcome (win, draw, loss) its the kind of thing that lends itself to being automated and learnt by a machine. Even checkers has been mastered by machine learning.

[Guybrush]

It definitely seems to me that many jobs may be replaced in part or entirely by AI / automation / robots. It's already happening, of course.

Here, I would like it if part of a solution would be shorter common workdays, so down to f.ex. 6 hours. If that causes a loss in productivity, businesses can make up for it by hiring more people.

It'd mean more strain on our HR department, but they deserve it anyways.

[Meatwad]

Id also rather it still be termed machine learning (the original way it was treated) and seen as elaborate code, rather than "AI" because it implies something it doesn't do which is form intelligent thought with its own autonomy. Its beholden to a set of inputs from the a specific user (who may or may not have a heavily biased opinion towards said input) to initiate whichever sequence or function its manifesting so its hardly "intelligent".

Additionally, I'd be concerned if I worked in a warehouse role for a major company that hasn't gone automated yet, so physical labouring jobs in warehouses should be replaced quite steadily over the next 5-10 years by the bigger operators. These types of investments that companies make to change to automation need their catchment population to justify the expense so it wont be a case of every company adopting it all of a sudden.

Coding will still continue to be a pretty reliable occupation going forward, but even the types of trades (like plumbing and electrical work) where automation would appear less likely to be utilised will be good options for those not entering tertiary education.

Heres an interesting video showing how labouring work has been replaced by robots in a grocery warehouse in Britain .........

[Marie Monday]

I have the same issue with the term 'artificial intelligence'. It's not limited to that; one of my issues with this branch of science is its use of misleading terminology that makes unwarranted implications

[Lisnaholic]

^ Sorry, Meatwad and Marie M; now that we have the handy acronym AI, that rolls off the tongue so easily, I doubt that anyone is going to change that label.

Having played chess quite a bit at an intermediate level, its actually not as unbelievable as people may think to map out the variables to it and design a program to play it at a strong level. Most of the opening play is learnt by rote and the endings have been studied and standardised many years ago. Coupled with middle game tenets of strategic play and feeding the program with tactical knowledge and you've essentially covered all the bases to play the game. Im sure if it was 1972 then these future developments that came along (internet, chess engines or otherwise) would have been more testing to conceptualise, but due to the nature of chess being a game played within a board with a fixed outcome (win, draw, loss) its the kind of thing that lends itself to being automated and learnt by a machine. Even checkers has been mastered by machine learning.

^ Yes, with the wisdom of hindsight, you make it sound like an easy logic progression to make a chess-playing computer. Back in 1972 the obstacle seemed to be the sheer number of permutations that were possible, and the way the options start accumulating: first move options, response options, second move options, etc. People thought that no machine could hold and juggle through all the possible moves.

I wonder how much chess you have played? I used to enjoy a game with my friends, but probably last played a game about 40 years ago. Me and my friends lost interest at about the time "serious" players began practicing like crazy with these machines:-

It'd mean more strain on our HR department, but they deserve it anyways.

^

Yep, reduced hours per week and job-sharing is one way to soften the blow of mass unemployment, but, alas, it's also costly and inefficient, so companies aren't likely to be keen.