Apparently, stealing other peopleās work to create product for money is now āfair useā as according to OpenAI because they are āinnovatingā (stealing). Yeah. Move fast and break things, huh?
āBecause copyright today covers virtually every sort of human expressionāincluding blogposts, photographs, forum posts, scraps of software code, and government documentsāit would be impossible to train todayās leading AI models without using copyrighted materials,ā wrote OpenAI in the House of Lords submission.
OpenAI claimed that the authors in that lawsuit āmisconceive[d] the scope of copyright, failing to take into account the limitations and exceptions (including fair use) that properly leave room for innovations like the large language models now at the forefront of artificial intelligence.ā
A comedian isnāt forming a sentence based on what the most probable word is going to appear after the previous one. This is such a bullshit argument that reduces human competency to āmonkey see thing to draw thingā and completely overlooks the craft and intent behind creative works. Do you know why ChatGPT uses certain words over others? Probability. It decided as a result of its training that one word would appear after the previous in certain contexts. It absolutely doesnāt take into account things like āmaybe this word would be better here because the sound and syllables maintains the flow of the sentenceā.
Baffling takes from people who donāt know what theyāre talking about.
I wish I could upvote this more than once.
What people always seem to miss is that a human doesnāt need billions of examples to be able to produce something thatās kind of āeh, close enoughā. Artists donāt look at billions of paintings. They look at a few, but do so deeply, absorbing not just the most likely distribution of brushstrokes, but why the painting looks the way it does. For a basis of comparison, I did an art and design course last year and looked at about 300 artworks in total (course requirement was 50-100). The research component on my design-related degree course is one page a week per module (so basically one example from the field the module is about, plus some analysis). The real bulk of the work humans do isnāt looking at billions of examples: itās looking at a few, and then practicing the skill and developing a process that allows them to convey the thing theyāre trying to express.
If the AI models were really doing exactly the same thing humans do, the models could be trained without any copyright infringement at all, because all of the public domain and creative commons content, plus maybe licencing a little more, would be more than enough.
Exactly! You can glean so much from a single work, not just about the work itself but who created it and what ideas were they trying to express and what does that tell us about the world they live in and how they see that world.
This doesnāt even touch the fact that Iām learning to draw not by looking at other drawings but what exactly Iām trying to draw. I know at a base level, a drawing is a series of shapes made by hand whether itās through a digital medium or traditional pen/pencil and paper. But the skill isnāt being able replicate other drawings, itās being able to convert something I can see into a drawing. If Iām drawing someone sitting in a wheelchair, then Iāll get the pose of them sitting in the wheelchair but I can add details I want to emphasise or remove details I donāt want. Thereās so much that goes into creative work and Iām tired of arguing with people who have no idea what it takes to produce creative works.
It seems that most of the people who think what humans and AIs do is the same thing are not actually creatives themselves. Their level of understanding of what it takes to draw goes no further than āwell anyone can draw, children do it all the timeā. They have the same respect for writing, of course, equating the ability to string words together to write an email, with the process it takes to write a brilliant novel or script. They donāt get it, and to an extent, thatās fine - not everybody needs to understand everything. But they should at least have the decency to listen to the people that do get it.
Well, thatās not me. Iām a creative, and I see deep parallels between how LLMs work and how my own mind works.
Either youāre vastly overestimating the degree of understanding and insight AIs possess, or youāre vastly underestimating your own capabilities. :)
This whole AI craze has just shown me that people are losing faith in their own abilities and their ability to learn things. Iāve heard so many who use AI to generate āartworkā argue that they tried to do art āfor yearsā without improving, and hence have come to conclusion that creativity is a talent that only some have, instead of a skill you can learn and hone. Just because they didnāt see results as fast as theyād have liked.
Very well said! Creativity is definitely a skill that requires work, and for which there are no short cuts. It seems to me that the vast majority of people using AI for artwork are just looking for a short cut, so they can get the results without having to work hard and practice. The one valid exception is when itās used by disabled people who have physical limitations on what they can do, which is a point thatās brought up occasionally - and if that was the one and only use-case for these models, I think a lot of artists would actually be fine with that.
I started drawing seriously when I was 14. Looking at my old artwork, I didnāt start improving fast until I was around 19 or 20. Not to say I didnāt improve at all during those five to six years but the pace did get faster once I had ālearned to learnā so to say. That is to say it can take a lot of patience to get to a point where you actually start seeing improvement fast enough to stay motivated. But it is 100% worth it because at the end you have a lot of things you have created with your own two hands.
And regarding the point on physical limitations, I canāt blame anyone in a situation like that for using AI if they have no other chance for realising their imaginations. For others, it is completely possible and not reserved for people who have some mythical innate talent. Just grab a pen or a brush and enjoy the process of honing a fine skill regardless of the end result. ā¤ļø
Alternatively, you might be vastly overestimating human āunderstanding and insightā, or how much of it is really needed to create stuff.
Average humans, sure, donāt have a lot of understanding and insight, and little is needed to be able to draw a doodle on some paper. But trained artists have a lot of it, because part of the process is learning to interpret artworks and work out why the artist used a particular composition or colour or object. To create really great art, you do actually need a lot of understanding and insight, because everything in your work will have been put there deliberately, not just to fill up space.
An AI doesnāt know why itās put an apple on the table rather than an orange, it just does it because human artists have done it - it doesnāt know what apples mean on a semiotic level to the human artist or the humans that look at the painting. But humans do understand what apples represent - they may not pick up on it consciously, but somewhere in the backs of their minds, theyāll see an apple in a painting and itāll make the painting mean something different than if the fruit had been an orange.
Interestingly, LLMs seem to show emerging semiotic organization. By analyzing the activation space of the neural network, related concepts seem to get trained into similar activation patterns, which is what allows LLMs to zero shot relationships when executed at a ātemperatureā (randomness level) in the right range.
Pairing an LLM with a stable diffusion model, allows the resulting AI toā¦ well, judge by yourself: https://llm-grounded-diffusion.github.io/
Children learn by watching others. We are trained from millions of examples starting from before birth.
When you look at one painting, is that the equivalent of one instance of the painting in the training data? There is an infinite amount of information in the painting, and each time you look you process more of that information.
Iād say any given painting you look at in a museum, you process at least a hundred mental images of aspects of it. A painting on your wall could be seen ten thousand times easily.
Thatās what humans do, though. Maybe not probability directly, but we all know that some words should be put in a certain order. We still operate within standard norms that apply to aparte group of people. LLMās just go about it in a different way, but they achieve the same general result. If Iām drawing a human, that means thereās a āhandā here, and a āheadā there. āHeadā is a weird combination of pixels that mostly look like this, āhandā looks kinda like that. All depends on how the model is structured, but tell me thatās not very similar to a simplified version of how humans operate.
Yeah but the difference is we still choose our words. We can still alter sentences on the fly. I can think of a sentence and understand verbs go after the subject but I still have the cognition to alter the sentence to have the effect I want. The thing lacking in LLMs is intent and Iām yet to see anyone tell me why a generative model decides to have more than 6 fingers. As humans we know hands generally have five fingers and thereās a group of people who donāt so unless we wanted to draw a person with a different number of fingers, we could. A generative art model canāt help itself from drawing multiple fingers because all it understands is that āfinger + finger = handā but it has no concept on when to stop.
And thatās the reason why LLM generated content isnāt considered creative.
I do believe that the person using the device has a right to copyright the unique method they used to generate the content, but the content itself isnāt anything worth protecting.
You say that yet I initially responded to someone who was comparing an LLM to what a comedian does.
There is no unique method because thereās hardly anything unique you can do. Two people using Stable Diffusion to produce an image are putting in the same amount of work. One might put more time into crafting the right prompt but thatās not work youāre doing.
If 90% of the work is handled by the model, and you just layer on whatever extra thing you wanted, that doesnāt mean you created the thing. That also implies you have much control over the output. Youāre effectively negotiating with this machine to produce what you want.
Wouldnāt that lead to the same argument as originally brought against photography, though?
A photographer is effectively negotiating with the sun, the sky and everything else to hopefully get the result they are looking for on their device.
One difference is that the photographer has to go the places theyāre taking pictures of.
Another is that photography isnāt comparable to paintings and it never has been. Iām willing to bet photography and paintings have never coexisted in a contest. Except, when people say their generative art is comparable to what artists have been producing by hand, they are admitting that generative art has more in common with photography than it does with hand-crafted art but they want the prestige and recognition those artists get for their work.
deleted by creator
Thats not work to you? My company pays me to spend time to do the right thing, even though most of the work does the computer.
I see where you are going at, but your argument also invalidates other forms of human interaction and creating.
In my country copyright can only be granted if a certain amount of (human) work went into something. Any work.
The difficult part is finding out whats enough and what kind of work qualify to lead to some kind of protection, even if partial.
The difficult part was not to create something, but to prove someone did or didnt put enough work into it.
I think we can hold generated or assisted goods to the same standard.
Putting a simple prompt together should probably not be granted protection as no significant work went into it. But refining it, editing the resultā¦ maybe thats enough, thats really up to the society to decide.
At the same time we have to balance the power of machines against human work, so the human work doesnt get totally invalidated, but rather shifted and treated as sub-type.
Machines already replaced alot of work, also creative ones. Book-printing, forging, producing foodā¦ the scary part about generative AI is mainly the speed of them spreading.
So as a data analyst a lot of my work is done through a computer but I can apply my same skills if someone hands me a piece of paper with data printed on it and told me to come up with solutions to the problems with it. I donāt need the computer to do what I need to do, it makes it easier to manipulate data but the degree of problem solving required needs to be done by a human and thatās why itās my job. If a machine could do it, then they would be doing it but they arenāt because contrary to what people believe about data analysis, you have to be somewhat creative to do it well.
Crafting a prompt is an exercise in trial and error. Itās work but itās not skilled work. It doesnāt take talent or practice to do. Despite the prompt, you are still at the mercy of the machine.
Even by the case youāve presented, I have to ask, at what point of a human editing the output of a generative model constitutes it being your own work and not the machineās? How much do you have to change? Can you give me a %?
Machines were intended to automate the tedious tasks that we all have to suffer to free up our brains for more engaging things which might include creative pursuits. Automation exists to make your life easier, not to rob you of lifeās pursuits or your livelihood. It never shouldāve been used to produce creative work and I find the attempts to equate this abominationās outputs to what artists have been doing for years, utterly deplorable.
I donāt choose my words man. I get a vague sense of the meaning I want to convey and the words just form themselves.
As an artist you draw with an understanding of the human body, though. An understanding current models donāt have because they arenāt actually intelligent.
Maybe when a human is an absolute beginner in drawing they will think about the different lines and replicate even how other people draw stuff that then looks like a hand.
But eventually they will realise (hopefully, otherwise they may get frustrated and stop drawing) that you need to understand the hand to draw one. Itās mass, itās concept or the idea of what a hand is.
This may sound very abstract and strange but creative expression is more complex than replicating what we have seen a million times. Itās a complex function unique to the human brain, an organ we donāt even scientifically understand yet.
Thatās not the point though. The point is that the human comedian and the AI both benefit from consuming creative works covered by copyright.
Yeah except a machine is owned by a company and doesnāt consume the same way. It breaks down copyrighted works into data points so it can find the best way of putting those data points together again. If you understand anything at all about how these models work, they do not consume media the same way we do. It is not an entity with a thought process or consciousness (despite the misleading marketing of āAIā would have you believe), itās an optimisation algorithm.
Itās a glorified autocomplete.
Itās so funny that this is something new. This was Grammarlyās whole schtick since before ChatGPT so how different is Grammarly AI?
Here is the bigger picture: The vast majority of tech illiterate people think something is AI because duh its called AI.
Its literally just the power of branding and marketing on the minds of poorly informed humans.
Unfortunately this is essentially a reverse Turing Test.
The vast majority of humans do not know anything about AI, and also a huge majority of them can also barely tell the difference between, currently in some but not all forms, output from what is basically a brute force total internet plagiarism and synthesis software, from many actual human created content in many cases.
To me this basically just means that about 99% of the time, most humans are actually literally NPCs, and they only do actual creative and unpredictable things very very rarely.
I call it AI because itās artificial and itās intelligent. Itās not that complicated.
The thing we have to remember is how scary and disruptive AI is. Given that fear, it is scary to acknowledge that we have AI emerging into our world. Because it is scary, that pushes us to want to ignore it.
Itās called denial, and itās the best explanation for why people arenāt willing to acknowledge that LLMs are AI.
It meets almost none of the conceptions of intelligence at all.
It is not capable of abstraction.
It is capable of brute force understanding similarities between various images and text, and then presenting a wide array of text and images containing elements that reasonably well emulate a wide array of descriptors.
This is convincing to many people that it has a large knowledge set.
But that is not abstraction.
It is not capable of logic.
It is only capable of again brute force analyzing an astounding amount of content and then producing essentially the consensus view on answers to common logical problems.
Ask it any complex logical question that has never been answered on the internet before and it will output irrelevant or inaccurate nonsense, likely just finding an answer to a similar but not identical question.
The same goes for reasoning, planning, critical thinking and problem solving.
If you ask it to do any of these things in a highly specific situation even giving it as much information as possible, if your situation is novel or even simply too complex, it will again just spit out a non sense answer that is basically going to be very inadequate and faulty because it will just draw elements together from the closest things it has been trained on, nearly certainly being contradictory or entirely dubious due to being unable to account for a particularly uncommon constraint, or constraints that are very uncommonly faced simultaneously.
It is not creative, in the sense of being able to generate something novel or new.
All it does is plagiarize elements of things that are popular and have many examples of and then attempt mix them together, but it will never generate a new art style or a new genre of music.
It does not even really infer things, is not really capable of inference.
It simply has a massive, astounding data set, and the ability to synthesize elements from this in a convincing way.
In conclusion, you have no idea what you are talking about, and you yourself literally are one of the people who has failed the reverse Turing Test, likely because you are not very well very versed in the technicals of how this stuff actually works, thus proving my point that you simply believe it is AI because of its branding, with no critical thought applied whatsoever.
Current models arenāt intelligent. Not even by the flimsy and unprecise definition of intelligence we currently have.
Wanted to post a whole rant but then saw vexikron already did so I spare you xD
deleted by creator
And human comedians regularly get called out when they outright steal others material and present it as their own.
The word for this is plagiarism.
And in OpenAIs framework, when used in a relevant commercial context, they are functionally operating and profiting off of the worlds most comprehensive plagiarism software.
They get called out when they use others work as a template, not as training data.
You do know that comedians are copying each others material all the time though? Either making the same joke, or slightly adapting it.
So in the context of copyright vs. model training i fail to see how the exact process of the model is relevant? At the end copyrighted material goes in and material based on that copyrighted material goes out.
you know how the neurons in our brain work, right?
because if not, well, itās pretty similarā¦ unless you say thereās a soul (in which case we canāt really have a conversation based on fact alone), weāre just big olā probability machines with tuned weights based on past experiences too
You are spitting out basic points and attempting to draw similarities because our brains are capable of something similar. The difference between what youāve said and what LLMs do is that we have experiences that we are able to glean a variety of information from. An LLM sees text and all itās designed to do is say āx is more likely to appear before y than zā. If you fed it nonsense, it would regurgitate nonsense. If you feed it text from racist sites, it will regurgitate that same language because thatās all it has seen.
Youāll read this and think āthatās what humans do too, right?ā Wrong. A human can be fed these things and still reject them. Someone else in this thread has made some good points regarding this but Iāll state them here as well. An LLM will tell you information but it has no cognition on what itās telling you. It has no idea that itās right or wrong, itās job is to convince you that itās right because thatās the success state. If you tell it itās wrong, thatās a failure state. The more you speak with it, the more fail states it accumulates and the more likely it is to cutoff communication because itās not reaching a success, itās not giving you what you want. The longer the conversation goes on, the more crazy LLMs get as well because itās too much to process at once, holding those contexts in its memory while trying to predict the next one. Our brains do this easily and so much more. To claim an LLM is intelligent is incredibly misguided, it is merely the imitation of intelligence.
but thatās just a matter of complexity, not fundamental difference. the way our brains work and the way an artificial neural network work arenāt that different; just that our brains are beyond many orders of magnitude bigger
thereās no particular reason why we canāt feed artificial neural networks an enormous amount of ā¦ letās say tangentially related experiential information ā¦ as well, but in order to be efficient and make them specialise in the things we want, we only feed them information thatās directly related to the specialty we want them to perform
thereās someā¦ āpre trainingā or āpre-existing stateā that exists with humans too that comes from genetics, but iād argue thatās as relevant to the actual task of learning, comprehension, and creating as a BIOS is to running an operating system (that is, a necessary precondition to ensure the correct functioning of our body with our brain, but not actually what youād call the main function)
iām also not claiming that an LLM is intelligent (or rather iād prefer to use the term self aware because intelligent is pretty nebulous); just that the structure it has isnāt that much different to our brains just on a level thatās so much smaller and so much more generic that you canāt expect it to perform as well as a human - you wouldnāt expect to cut out 99% of a humans brain and have them be able to continue to function at the same level either
i guess the core of what iām getting at is that the self awareness that humans have is definitely not present in an LLM, however i donāt think that self-awareness is necessarily a pre-requisite for most things that we call creativity. i think thatās itās entirely possible for an artificial neural net thatās fundamentally the same technology that we use today to be able to ingest the same data that a human would from birth, and to have very similar outcomesā¦ given that belief (and iām very aware that it certainly is just a belief - we arenāt close to understanding our brains, but i donāt fundamentally thing thereās anything other then neurons firing that results in the human condition), just because you simplify and specialise the input data doesnāt mean that the process is different. you could argue that itās lesser, for sure, but to rule out that it can create a legitimately new work is definitely premature
āSoulā is the word we use for something we donāt scientifically understand yet. Unless you did discover how human brains work, in that case I congratulate you on your Nobel prize.
You can abstract a complex concept so much it becomes wrong. And abstracting how the brain works to āitās a probability machineā definitely is a wrong description. Especially when you want to use it as an argument of similarity to other probability machines.
thatās far from definitive. another definition is
but since we arenāt arguing semantics, it doesnāt really matter exactly, other than the fact that itās important to remember that just because you have an experience, belief, or view doesnāt make it the only truth
of course i didnāt discover categorically how the human brain works in its entirety, however most scientists iām sure would agree that the method by which the brain performs its functions is by neurons firing. if you disagree with that statement, the burden of proof is on you. the part we donāt understand is how it all connects up - the emergent behaviour. we understand the basics; thatās not in question, and you seem to be questioning it
itās not abstracted; itās simplifiedā¦ if what youāre saying were true, then simplifying complex organisms down to a petri dish for research would be āabstractedā so much it ābecomes wrongā, which is categorically untrueā¦ itās an incomplete picture, but that doesnāt make it either wrong or abstract
*edit: sorry, it was another comment where i specifically said belief; the comment you replied to didnāt state that, however most of this still applies regardless
i laid out an a leads to b leads to c and stated that itās simply a belief, however itās a belief thatās based in logic and simplified concepts. if you want to disagree thatās fine but donāt act like you have some āevidenceā or āproofā to back up your claimsā¦ all weāre talking about here is belief, because we simply donāt know - neither you nor i
and given that all of this is based on belief rather than proof, the only thing that matters is what we as individuals believe about the input and output data (because the bit in the middle has no definitive proof either way)
if a human consumes media and writes something and it looks different, thatās not a violation
if a machine consumes media and writes something and it looks different, youāre arguing that is a violation
the only difference here is your belief that a human brain somehow has something āmoreā than a probabilistic model going onā¦ but again, thatās far from certain
Am I a moron? How do you have more upvotes than the parent comment, is it because youāre being more aggressive with your statement? I feel like you didnāt quite refute what the parent comment said. Youāre just explaining how Chat GPT works, but youāre not really saying how it shouldnāt use our established media (copyrighted material) as a reference.
I donāt control the upvotes so I donāt know why thatās directed at me.
The refutation was based on around a misunderstanding of how LLMs generate their outputs and how the training data assists the LLM in doing what it does. The article itself tells you ChatGPT was trained off of copyrighted material they were not licensed for. The person I responded to suggested that comedians do this with their work but thatās equating the process an LLM uses when producing an output to a comedian writing jokes.
Edit: Apologies if I do come across aggressive. Since the plagiarism machine has been in full swing, the whole discourse around it has gotten on my nerves. Iām a creative person, Iāve written poems and short stories, Iām writing a novel and I also do programming and a whole host of hobbies so when LLMs are used to put people like me out of a job using my own work, why wouldnāt that make me angry? What makes it worse is that Iām having to explain concepts to people regarding LLMs that they continue to defend. I canāt stand it so yes, I will come off aggressive.
Sorry, I was essentially emphasizing on my initial point āam I a moron?ā, lol, because I legitimately didnāt get your point at first like others do in this thread.
I get what you mean now after reading it couple more times
Text prediction seems to be sufficient to explain all verbal communication to me. Until someone comes up with a use case that humans can do that LLMs cannot, and I mean a specific use case not general high level concepts, Iām going to assume human verbal cognition works the same was as an LLM.
We are absolutely basing our responses on what words are likely to follow which other ones. Itās literally how a baby learns language from those around them.
If you ask an LLM to help you with a legal brief, itāll come up with a bunch of stuff for you, and some of it might even be right. But itāll very likely do things like make up a case that doesnāt exist, or misrepresent a real case, and as has happened multiple times now, if you submit that work to a judge without a real lawyer checking it first, youāre going to have a bad time.
Thereās a reason LLMs make stuff up like that, and itās because they have been very, very narrowly trained when compared to a human. The training process is almost entirely getting good at predicting what words follow what other words, but humans get that and so much more. Babies arenāt just associating the sounds they hear, theyāre also associating the things they see, the things they feel, and the signals their body is sending them. Babies are highly motivated to learn and predict the behavior of the humans around them, and as they get older and more advanced, they get rewarded for creating accurate models of the mental state of others, mastering abstract concepts, and doing things like make art or sing songs. Their brains are many times bigger than even the biggest LLM, their initial state has been primed for success by millions of years of evolution, and the training set is every moment of human life.
LLMs arenāt nearly at that level. Thatās not to say what they do isnāt impressive, because it really is. They can also synthesize unrelated concepts together in a stunningly human way, even things that theyāve never been trained on specifically. Theyāve picked up a lot of surprising nuance just from the text theyāve been fed, and itās convincing enough to think that something magical is going on. But ultimately, theyāve been optimized to predict words, and thatās what theyāre good at, and although theyāve clearly developed some impressive skills to accomplish that task, itās not even close to human level. They spit out a bunch of nonsense when what they should be saying is āI have no idea how to write a legal document, you need a lawyer for thatā, but that would require them to have a sense of their own capabilities, a sense of what they know and why they know it and where it all came from, knowledge of the consequences of their actions and a desire to avoid causing harm, and they donāt have that. And how could they? Their training didnāt include any of that, it was mostly about words.
One of the reasons LLMs seem so impressive is that human words are a reflection of the rich inner life of the person youāre talking to. You say something to a person, and your ideas are broken down and manipulated in an abstract manner in their head, then turned back into words forming a response which they say back to you. LLMs are piggybacking off of that a bit, by getting good at mimicking language they are able to hide that their heads are relatively empty. Spitting out a statistically likely answer to the question āas an AI, do you want to take over the world?ā is very different from considering the ideas, forming an opinion about them, and responding with that opinion. LLMs arenāt just doing statistics, but you donāt have to go too far down that spectrum before the answers start seeming thoughtful.