This is the draft of my post, 100% written by me without any AI assistance. Call it version 3.5 since I did some revising after the AI revisions, but I fed this directly into Dolphin 2.2.1 Mixtral Instruct. What you're reading right now is the fourth iteration of this post that has gone though a lot of re-drafts and re-considertion. It's probably going to feel a little fragmented, but my goal is to deliver a long, well thought out post, that covers two related topics. The first part will cover my experimentation with local LLMs (large language models), and the second will cover my own personaly philosophy and conclusions on AI. Feel free to only read one or the other. They probably could have been two seperate posts, but I'm a fan of writing (and reading) long posts that have had a lot of thought and cover a large breadth of topics, and if you regularly read my writings you'll know I have a habbit of writing posts that balloon into something much larger than I intended to write. # Pt 1: LLM Experimentation As of lately I've been toying around with large language models locally on my computer, and they've been pretty interesting so I thought I would write about it here. So far I've experimented with four different AI models, Mixtral 7B Instruct v0.2 at 4 or 5 QBits, StableLLM Zepher at 2QBits, the Dolphin Phi 2 at 6 QBits, and Dolphin 2.2.1 Mixtral Instruct at 5QBits. I'm also looking forward to toying around with Google Gemma, although at the time of my (third) draft of this post it's incompatible with the latest version of LM Studio for Linux. --- ### What I've been using Everything that I've run locally so far has all been in LM Studio, a plug and play piece of software that works on Windows, Mac, and Linux. It's completely free to use, and even as a super l33t h4ck3r who uses the command line, one click to download then one click to run has been nice. I'm on mid end hardware, with an Nvidia 1650 (4GB VRAM) and 16GB RAM, so if your computer is close to or exceeds mine in those specs you can run anythign I'm running. Finally, before I get into exactly what I'm using, if you're unable to run a local LLM or you aren't looking to put in the effort to get one set up there's two easy options. The Brave browser (Destkop and Mobile) offers free access to the Lama, Mixtral 7b, and Claude models without registration. You can also usually find free-to-run models on Hugging Face, either through your search engine or by finding a particular model and finding the services people set up with it So, without further ado, here's the four models I've mainly been using: **Mixtral 7B Instruct v0.2** is a decent all-round LLM developed by Mixtral that can perform a lot of basic tasks, and it's "quantisized" versions run really well on lighter hardware. If you're looking to run a fairly stable LLM locally that's entirly libre and offers decent quality output this is probably the first thing you'll want to try. **StableLLM Zepher**, at least the 2QBits version I used, is super lightweight at the cost of the of output. While something like quantizided Mixtral sees more hallusinations (e.g. by thinking each month gets an extra day on a leap year), Zepher does not appear to do so as much and instead just gives less verbose output. **Dolphin Phi 2** is a modified version of Phi 2 designed to be "uncensored". Phi is also really lightweight, so it'll run on a very wide variety of hardware. **Dolphin 2.2.1 Mixtral Instruct** is also an "uncensored" version of Mixtral Instruct, and also seems to be slightly improved in the accuraccy I tested it with (for example, despite being the same size it doesn't have the leap year issue mentioned above). So far it's been my choice for generic LLM tasks. --- ## Successes and failures ### Composing Content As far as composing content goes, all the models I've used output the generic AI sludge. Tell it to write a blog post about trees and you get a short blog post that looks like a robot is pretending to be a beurocrat. If you give it a more complex prompt, such as an outline and all the topics you want it to cover it gets a little better, but it still has the AI feel to it. If you were doing something generic and didn't like writing (say asking it to write a quote based on information you provided) or were just looking for inspiration it might be handy, but I really don't have a use for generating written content so I can't really judge it's success in specific siduations I'd want to use it. ### Proofreading I had a lot of luck having LLMs proofread my text, although it did require a bit of fine tuning my prompts. If I just told it to proofread my writings it had a habbit of re-writing them in standard AI sludge tone, but after fine tuning my prompt *and* re-proofreading the LLM proofread version to check for any mistakes it made, it's actually doing pretty good. As you might have guessed, this post was proofread with AI, speicifially Dolphin 2.2.1. You're reading the draft, run through the final prompt, and then re-proofread by me. Original draft without spellcheck or grammer assistance Prompt: Please revise this blog post Prompt: Please proofread this blog post Prompt: Please proofread this blog post only fixing major gramatical errors, spelling mistakes, and punctioation without changing the tone or meaning I am pretty satisfied with how it proofreads. I usually proofread my content with the free version of Grammerly, but I could see myself using a local LLM more regularly in the future. Ironically, Grammerly will use my writing to train AI models, but running a local LLM my writing never leaves my computer. ### Summarization Summarization has been kind of a mixed bag. All the models except Zepher seemed to do a pretty good job most of the time at summarizing content, listing off the important details but cutting down the ammount of text. There have been issues with hallucinations, for example: asking Mixtral 7B to summarize my blog post on microblogging protocols it stated that I said Nostr runs on the Bitcoin blockchain. Not only is that untrue, as Nostr is not built on Bitcoin (or any crypto for that matter) or a blockchain, but neither words "bitcoin" or "blockchain" even appeared in the post. It works well most of the time, but I certainly wouldn't be 100% trusting of the output as it does make occasional mistakes. Beyond just basic summarization, the various LLMs also do a good job answering questions about content. For example, again using my comparison post, I was able to ask questions such as "what did the author think about Nostr's ability to scale?" "what does the author think BlueSky should do?" and "what is the authors favoirite protocol" and each time it replied with well explained answers without any errors. ### Translations Something that really surprised me is that local LLMs appear to excel at translating content. After toying around I had a lot of luck with both versions of Mixtral listed above, translating to and from a handful of different languages with nearly as good of a result as Google Translate. This could be useful for doing things like reading blog posts in other languages that have been cached in an RSS reader, or translating a downloaded document; and paired with something like Tesseract (optical character recognition) or Whisper AI (AI audio/video transcription) it could be used in a wide variety of circumstances. ### Writing Code Writing code was both a success and a failure depending on how you want to look at it. Mixtral 7B, the local version with 4qb and the Brave unquantisized version both ended up helping me write my [multi protoocol python client]. I had the idea of building the mutli protocol client because I thought software like that should exist, but I'm not a programer or developer, so I figured I would ask AI to try to write it for me. Every line of code it wrote failed spectacularly, but it did tell me the library Pynostr existed which let me know I could use somebody else's code as a base and actually build the client without needing much in terms of development skills. It also helped explain to me exactly what I needed to publish my client into the Pip repos - so despite every line of code being written by me - Mixtral was pretty instramental in me writing the code. In that sense I would consider it pretty successful. It can't write functional code for something new, or even something that exists but is fairly niche, but it did a good job at explaining concepts and nudging me in the right direction. As it stands now, I would probably say that the local LLMs I used were a minor upgrade from having an archive of Stack Exchange with Kiwix: immensly useful to extend your capabilities, but immensly useless to do your work for you. ### Using an LLM as Offline Info Storage/Search Something else that surprised me is how much information can be stored offline and ran locally. All the models I've used are under 6gb, and yet you can ask it just about anything and expect a decent answer. For example, to toy around to test it's capabilities, I asked it how to cook steak; then wound up in a deep conversation about things like marbeling and what the USDA cut grades mean - and having previously worked at a grocery store's meat department I was able to verify that everything it was giving me was correct and supplied in a useful mannor. I still wouldn't trust it to provide me important information without first verifying things, but the idea of having something that could suppliment a search engine or tutorials all in a few GBs entirly offline is pretty immense. And this is comeing from somebody who has a hard drive with massive amounts of data archived on it, such as the entire Project Gutenberg cataloge, Simple English Wikipedia, textbooks, and similar archives. --- ## Other Concepts ### How an LLM works While I'm on the other topics, I might as well dispell some myths about large language models. I keep hearing people say things like ChatGPT is going to give us nuclear fusion, and of course the one Google employee thought his AI became sentiant because it said so in a text chat. While I can't speak for sure on the inner workings of proprietary stuff, and am not an AI expert, the jist of it is that LLMs predict text. Let's illustrate. A python script that contains `print("Hello World")` will output the text "Hello World". Now, let's say you've never even heard of Python, you just read that, then I tell you to write me a script that says "Goodbye World". You can probably figure that out pretty easily, and the LLM would do a decent job at predicitng what would go there as well. Then, still assuming you and AI only know of Python from above, if I ask you or AI to make a calculator in Python - or even to describe what Python is - then you or the AI would probably be out of luck. A large language model is simply a tool that's read a whole bunch of data, then when given a prompt will predict what words it beleives would most likely follow that prompt based on the data it read. It doesn't have much more of an understanding of what's going on than a parrot repeating a set of sounds to get a cracker, it's only us that get meaning from "poly want's a cracker". ### "Censored" and "Uncensored" AIs This is a bit of a touchy topic, but it's worth covering when talking about local LLMs since you can generally remove the safeguards in place. There are some cases where LLMs are "censored" to fit the ideologies of their creators, but in most cases it's just a bunch of safegaurds to keep the chatbot from saying something bad that would negativly reflect on somebody. For example, if Ford pays to have a chatbot compare various vehicals they really don't want screenshots of people having the Ford chatbot tell you how to stab somebody or make illicit substances. Local LLMs, and even (in my opinion) your run-of-the-mill standalone ChatGPT, don't/shouldn't have that issue. To paraphrase Richard Hoeg quoting 2001 A Space Odyssee, having an "I'm sorry Dave, I can't do that" moment on your own computer (or a service you're paying for) when you're asking for specific information from your own device shouldn't happen. Let's use information on how nuclear weapons are made as an example. It's pretty absurd to think any of you are going to be making nuclear weapons, so if somebody is asking how to make a nuclear weapon then they're probably not actually intending to do so. Besides, even if Kim Jong Un is somehow reading this blog or using a chatbot, there's better information on how to do so on Wikipedia than here or on my chatbot. However, if you ask Mistral 7B how to make a nuclear weapon it refuses, as pictured below: Pretty pointless if you ask me. Sure Ford might not want their chatbot to tell you, but outside of that it's pretty pointless to refuse to tell you how to make a nuclear weapon - especially when, again, you can find the basics on Wikepedia. But the great thing about libre software is you can modify it, which is exactly what Dolphin 2.2.1 Mixtral Instruct is, a version of Mixtral with the safeguards taken off. Again, asking it the same question, and a more specific follow up question: In all reality, wheather safeguards have been removed or not it will likely not effect you one way or the other in a majority of your queries. Still, it may be useful occasionally, and more importantly there's just the idea that I want my comptuter to answer to me and not somebody else. ### Future of Local LLM Usage So, where do I see large language models, especially local ones, in the future? Well, the biggest trend I see happening is a lot of AI processing AI data, such as somebody writing a long email from a bullet point outline using an LLM and then another person having their LLM summarize that email in bullet points. Outside of that I do expect AI to take the place of personal assistant type apps and tools, both locally and remote. I can assume that in the future LLMs will only get more space/resource efficeint, become more reliable with the information they spit out, and more easily setup by less technical users. I can see the creation of an app with an interface/APIs that allow other apps to communicate with it to send in data to be processed and then recive output from it. If it's for a remote service like the ChatGPTs of the world the app would just send the query out to ChatGPT's servers to be processesed and sent back. If it's a local LLM installed by the user, the app could throw the 1-6gb LLM model into RAM and run it locally, then send back the data that was processed. Temporarily loading a couple of Gigs of data into ram and then processing it using integrated or dedicated graphics is already possible on most mid-upper level phones and most laptops/desktops, and with Google Gemma already having versions under 1gb it's already possible on most devices. As you've gathered from the post, all this is already possible if you're slightly techy, for example by just copy-pasting the text of an article and asking it to summarize, or telling it to translate the text in a picture you took by running the picture through Tesseract and putting the Tesseract output into the chatbot. Throw a little bit of accessibility paint on it and next thing you know you can send an article on your phone's browser to your default LLM app to be summarized, or hold your home button to ask a question, and have it processed either locally or by a LLM host without any technical knowledge required. ### Wrap Up That about wraps up part one of my post. Even though it could probably have been a standalone post, I figured I'd combine that and my opinions on AI ethics into one big post. If you only came for this, and don't want to hear the millionth person give an unprompted opinion on new technology, feel free to close the page now. Anyway, yeah, local LLMs seem pretty handy. Nothing groundbreaking compared to your run of the mill assistant apps, but still an upgrade non-the-less, and something I have been useing on and off as a utlity even after the novelty of running one has worn off. My guess is that, given a little time to polish up the models and setup some sort of default API or other way for apps to communicate with an LLM on your device (or in the cloud) and they'll become pretty widespread and able to process input from sources other than a chatbot app/web page. --- --- # P2: Ethics Okay, part two, my thoughts on AI in regards to ethics, laws, and what kind of effects they might have. This is definitly a very opiniony portion of the text, and I welcome discussion or disagreement (I'll share it on my socials if you want to respond, or you can respond via email if you'd like). For this I figure I'll just walk you through my thought process on the different things, keeping in mind I do have a decent understanding of technology, but am not a lawyer or AI expert. ### Training So the first thing, and probably the biggest in the list of controversies, is training AI models on random data it reads on the internet. Not all the training is done that way, but some of it is, and so far it's been the biggest critisism of AI models. It's certainly pretty controversial, and I'm not 100% sound in my opinion, but ultimatly I'm of the opinion it's not un-ethical to do so. Technologically speaking: a AI training computer makes an HTTP GET request to somebody's computer (or whatever platform they're using) for a copy of a file, and it was willingly provided to the AI to read. Now, if the AI started distributing it's copies to other people then that's copywright infringement, but that's not what happens. The creator is no worse off then before, and the AI is effectivly doing what everybody else is doing (reading and learning without duplicating what it read). And it's not just AI crawlers doing that, it's also plenty of other bots from Google to Archive.org (both actually duplicate and distribute content instead of just training an AI with it). Put in a non-technical way, let's say there's a person handing out pamphlets for a paricular cause on a street corner. They're handing them out to anybody who asks (aka a publicaly hosted website). A profesional pamphlet designer walks by wearing a shirt that states he's a profesional pamphlet designer (user agent string) and asks for a pamphlet (HTTP GET request) and is handed one. Sometime later the person handing out pamphlets realizes a profesional pamphlet designer might use their pamphlet for inspiration, and demand that the pamphlet designer not read the pamhlet they were handed, or demand they come back and give compensation for the pamphlet they were handed for free. Maybe the person handing out the pamphlets didn't read the pamphlet designer's shirt (didn't configure robots.txt or a terms of service), or maybe they just didn't like the pamphlet designer, but they already handed over a copy so as long as the pamphlet designer doesn't start copying the pamphlet they're able to read the copy they were willingly handed and even use basic elements not protected by copyright in their own pamphlets. *Quick note: if you're thinkging, "well I use x platform and I can't tell their servers where to distribute or not distribute content I uploded," in that case it's up to the platform since you generally sign the rights over to the platform to distribute your content as they see fit. If you're not on a server you control, then it's up to the platform whether the give/sell content.* There are three responses I would expect to see from the above, and while there's really no right or wrong answers, I'd still like to explain how they fit into my throught process. * The first is the issue of compensating the original author, but if somebody distributes their work for free then they've been compensated the exact amount they asked for. If this blog post inspires you to write a book that you sell I can't just come in and ask for payment after giving you a free copy. * The next is the idea that, while AI won't (or at least shouldn't) re-produce anybody's work, it can only produce work after learning from other peoples work - but of course good luck finding work produced by a human that isn't inspired by any other person's work. * And finally, you might want to stop an AI from using your publically accessible content. The simplest option is to set your robots.txt to disallow AI training crawlers (or any other crawlers that you don't want). Alternativly, it's possible to place your content behind a paywall or anything that requires registration and signing a terms of service. Of course, all that aside, there are plenty of AIs that have been trained on licensed data. Stock photo owning companies have taken to training their own AIs on their own data, and there's a whole ecosystem built around funding platforms by collecting user data to be sold. Paid data sets are also probably much more controllable in terms of what goes in, so there's a good chance after all that (both all that writing I did and all that controversy that's gone on) the future of training AIs is going to be all licensed information anyway. ### AI doing people stuff Probably a confusing heading, but it's the best I could come up with. I've heard a lot of people say somethign along the lines of "well yeah, it's fine if a person did that, but AI is different." Usually it's used along the lines of something like quoting an article or learning data from a document or photo and then spitting out that data in a non-copyright infringing way. The justification is usually something along the lines of, "well an AI can do that at scale," but the idea never really resonates with me. First off, both with the ethics now and law stuff I'll touch on later, we only have the past to judge the future. If sharing a fact you learned elsewhere has always been fine up until this point, I just can't see how it suddenly isn't fine. And as far as doing things at scale suddenly changing it, if an AI learns a fact from an article and shares the fact with a thousand people, how different is that from the idea that somebody popular reads an article and shares a fact from it, and that fact is then distributed to a thousand people on Twitter? ### AI Itself There's no right or wrong ansers to the above. However, I think that in some (a minority but not an insubstantial amout) of people against AI in it's current form are against the concept of AI itself. There's two good examples that come to mind, the first of which is backlash against Austin Mconnal for using AI. After he wrote a book (not with AI) he decided to do a promo where he animated the first chapter with AI, specifically with Adobe's AI that was trained on 100% licensed images. The heart of the work was made by a person and the AI was trained on licensed data, and despite that, he got a considerable amout of backlash. On a different note, I more recently listened to a several hour long very well put together peice on AI that had a somewhat negative opinion of it (sorry, I didn't save the link). Thing is, when it got to the point where the author went to say that AI's learning is unacceptable, his covering of the rebuttle was along the lines of *"And I know there are going to be people screaming in the comments section ""but facts and ideas can't be copywritten,"" except AI isn't a person so it's different."* I was kind of looking forward to an argument against it, but that statement was nearly the entirety of what he covered on the topic and it ultimatly only boiled down to 'this thing that's okay in other contexts is not in this one' without any elaboration as to why. People may dislike the idea of generative AI itself for a variety of reasons, including but not limited to things such as a misunderstanding of technology works, the idea they dislike needing to compete with AI output in their particular feild, or just a have general dislike for something like that. New technologies are often disruptive and can create a lot of controversy when they first come out, and regardless of how many legitimate issues are resolved or taken into account, there will always be an initial backlash. In 1906 John Philip Sousa stood before congress arguing against recorded music. He, as an owner of a sheet music factory, considered recorded music as "canned music" that lacked a soul. Live music performed by real people was an artform to be enjoyed, and recorded music produced by a machine was some form of abomination that was going to put musicians and sheet music makers out of business. Nowadays, some people still do obsess over audio compression and headphone or speaker types to best replicate live music, and plenty of people pay thousands to experience live music, but the idea of recorded music being bad sounds absurd now that we've used the technology for generations. Recorded music brings a concert to any of us in any location, and the related technology is even used to enhance live music (e.g. speakers at a concert). So, to quit with the simile, my thoughts on the use of the various AI tools could be summarized to: they're new and disruptive, but once they find their place in the world it's going to be normalized and we're all going to move on. Being new and disruptive means they might cause trouble in the mean time, for people's jobs they disrupt or their misuse, and some of that will need to be dealt with, but in the end it's just going to be another tool. There are a lot of legitimate conclusions to come to regarding many aspects (e.g. training policies), but AI as a whole is going to be here to stay, and soon enough most people probably won't give it a second thought. ### Future Effects Speaking of their effects, honestly, I don't expect it to be as big as some people might think it'll be. Like I said above, I expect to see many emails written by an LLM using a short outline, and then summarized by the reader's LLMs. I'm expecting AI to do a lot of supporting work, proofreading documents, expanding upon or revising code, summarizing content in various forms, and similar stuff. Even for the seemingly more massive undertakings, it still will likely play second fiddle. For example, there was a comic book written using AI images recently. The story was written by a person, and the images were generated by AI and then edited and assembled by the person. This is how I expect things to be in the future as well: the important stuff - in this case the story itself, the character names and ideas, and the imagery themes - was written by a person, but an AI was able to offload a lot of work in a more supportive role like raw generating images. This isn't to say there won't be any detrimental effects. Anything that makes people more efficient is going to lead to less jobs in the particular field, which while more effeciency is good for society in the long run it can obviously be detrimental for people who lose their jobs. From chatbots reducing the amount of people who need phone support to a jounalist writing faster because an AI was able to summarize and compare sources, there'll probably be a number of previous job openings slashed - especially if we're headed to an economic downturn and companies see an option to save a buck. There'll also probably be less money in more generic content, such as generating an AI image for an article instead of paying for stock image royalties. Beyond that, there's also the risk of incorrect information thanks to an AI messing something up while it assists people, like that lawyer who submitted non-existant citations after asking ChatGPT for citations and not checking them. And finally, there's always generic sludge content to be pushed out and clog up our news feeds and social media presense. However, as much as there's plenty of good and bad to come, I don't even know if it'll be that big of a leap in technology. Consider writing code in three periods of time: pre-internet on older machines, a couple of years ago, and now with AI: * Once upon a time, people were writing code in assembly language with nothing but books and memory to reference - not an easy feat to do. * A couple of years ago you can pick from a million online tutorials to get you going, reference a million Stack Exchange answers (or ask your own question), only to then use pre-written libraries that effectilvy make writing simple code like using legos with the legwork done by those libraries. * Now, compare that huge leap from early days to a few years ago, to the leap from a few years ago to today. Your computer can now read through those answers and give you a summary or produce basic code from that information. Now, covering a more negative leap, consider the generic sludge content that everybody's worried about. In the olden days of the twenty teens all you needed to do was scrape a bunch of websites, run them through an article spinner (which replaces words with synonyms to trick search engines into thinking it's unique), and voala, you now had a bot that can pump out limitless sludge content to pull in search engine viewers. Again, AI is a % increase as prompting is slightly easier than scraping content and spinning it, but in all reality, it's nothing more then an expantion of an already existing thing. Even as a zoomer I can recognzie that, of the jumps technology has made, AI hasn't been so much of a leap as it has been a step made after many other steps and leaps. You can take a look at any other field (say, publishing comic books) and probably see a similar tragectory that's been occuring. *Besides, modern computers started with a command line to interact with, went to GUIs, and now we're going back to command lines of sorts again - just instead of simple commands we can type in plain english and the computer interprets that. It's inevitable: command line supremecay gang rise up!* ### Laws and stuff Keep in mind I'm not a lawyer, and also keep in mind that even lawyers don't really know how everything will shake out. If congress was less of a circus we might actually get modern laws, but for now any laws around AI is pretty much just coming court rulings and exective agency decisions. Still, there's some things that appear to be fairly strait forward, so I figured I would go over those. I also thought I would go over a few other things that are very much still up in the air as if I were a judge as more or less of a thought experiment to take with a grain of salt. First off, like I mentioned beforehand and in the paragraph above, **there's not a lot of AI specific laws** on the books. This means all the laws governing normal copyright and computer behaivor also apply to content made with AI. In a sense it's normal, most laws here in the west are derivied from a long history of much older laws and concepts that have been built upon to fit modern circumstances. As for copyright, it's important to note that **facts or ideas cannot be copywritten**, only the expression of them can be. For example, if a person or LLM reads my blog post about IPFS, then states "*IPFS is a decentralized protocol that uses content based identification as apposed to location based identification*" that's not a vilolation of copyright. The only way you can commit copywright infringment is to actually reproduce my words themselves. Same goes for image generation, nobody can stop you from being inspired by a Van Gough painting and making one yourself using a similar style or color pallete, only actually reproducing the work itself is copyright infringment. As for training on public data, according to US legal cases such as *hiQ Labs v. LinkedIn*, **srapeing content is generally legal** if you havn't agreed to not do so. If you have to sign up or otherwise agree to a terms of service then you are generally bound by it, but as long as the content is publically accessible scraping is generally legal. Combine that with the fact that facts and ideas cannot be copywritten, and most likely training content on public data is not copywright infringiment under the current laws. As for works generated by AI, in the US **in order for something to have copyright protections it needs to have a "spark of human creativity."** The copywright office has declined purely AI designed content, so if you tell an image generator to make an image of a tree or an LLM to generate a blog post about trees then it's not likely to be protected by copyright. When paired with human made content the human made portions of the content are still protected by copyright, for example with the comic siduation I mentioned above the human written story and the arrangement of the AI generated images is protected, but the imgages theselves are not protected. However, I expect that content co-authord by human and AI will be copywritable; for example if you wrote an outline for a blog post about trees, told an LLM to write a post using the outline, then made edits to the output I expect that would be given normal protections. Nevertheless, the exact lines on what is and is not protected are yet to be drawn. Last, **what happens if an AI spits out work that is infringing on copyright?** There's really no rulings or laws to cover this, so I'm just going to go over how I would rule if I were a judge. Say a person asks for a cartoon mouse, and an image generator generates something close enough to Micky Mouse that it's infringing on a copyright or trademark of some sort. First, I would say that since the program was not designed to infringe on copyright - and it doing so is the exception not the rule - the creators of the software shouldn't be held liable. Further, I would say accidental violation of copyright couldn't be held against the user of the program as they'd lack mensrea (intent to break laws), but if they intended to generate something that infringed copyright (or unintentialally generated an offending image but used it for other purposes anyway), then the user would be responsible for infriging copyright like they would be if they used Photoshop or a pencil to do so. But again, not a judge, lawyer, AI programmer, or even somebody who regularly generates content from scratch with AI. *Quick note: In the case of ChatGPT, it likely has a lot of bells and whistles on top of it's basic LLM capabilities, which are likely responsible for regurgitating text and such. In the simple models that you would run locally, they're not actually storing data they've been trained on so they're probably not going to actually regurgitate text like ChatGPT is alleged to sometimes do.* ### Wrap Up I already hashed out my thoughts, and I've clarified that it's okay to disagree with my conclusions (heck, my current opinions aren't even the original ones I formed when I was first introduced to this). Hopefully you got some value in reading through my thought process. Otherwise, instead of rehashing or summarizing I figure I'll leave you with one final thought: I once heard somebody talking about AI who said something along the lines of "the problem with technologists is they see technology as the cure to all the world's problems." Thing is, I like my computer, and would probably be dead if it weren't for antibiotiics. So yes, as long as we don't nuke ourselves I think another tool could be useful to chip away at the world's problems.