42
votes
Creatives, how do you feel about the impact of artificial intelligence on the future of art, illustration and design?
I will be participating in a panel discussion about the intersection of art and Artificial intelligence next week, and I am curious how fellow creatives feel about Artificial intelligence.
Have you used AI before in the creative process? If so, what services have you used/prefer?
What do you think the role of AI is in the creative process?
Does AI enhance creativity or limit originality?
What are the ethical implications of using AI to create art?
I'm a writer who has been using AI art to create visual media to accompany my writing. I think the largest shift that AI image generation represents is that it removes the step where you need to explain to another person what it is that you want. "Explaining" in a prompt is still something of a challenge, but the lack of another person in the iterative loop means that you are the one directing the way things change rather than trying to make yourself understood by another person.
It's obviously not as good as if I were able to draw, and I would never think of using it for something front and center like a book cover - those, I still commission. But for the art that would just be missing if I didn't have a nearly-free way to generate it, that doesn't involve convincing another person to spend hours of their time every week on my projects (even with pay, a dicey prospect) - it's definitely a game-changer.
The most obvious limitation is the lack of ability to specify detail, which I can get around with some freehand alterations and composition for the moment; with PS introducing inpainting and more robust tools available every month I imagine that will be a diminishing issue. Some of the advanced character and pose generation tools are pretty remarkable even now.
Concerning ethics - I'm of the opinion that AI will hit certain industries with all the subtle care of a nuke. Commission artwork, particularly the boilerplate corporate crap that is bread and butter for a lot of people, is going out the window. Writing-focused industries and administrative jobs will also be pretty hard-hit by this. Knowing that the technology is trained on a vast body of work that plausibly includes work by the very people the tech will replace is galling, and I understand the outrage some people are vocally expressing about AI.
That said, I think that calling AI inherently infringing simply as a consequence of the training is incorrect. Training builds a weighted model, which holds none of the work used to train it. Simplifications of the technology as "collaging" or copy/paste of existing material betray a lack of understanding of the tech. And while it certainly can be used (and will, inevitably, be used) to create infringing content, so too can copiers, Photoshop, etc. AI is infringing if you use it to infringe on copyright, just like any other creative tool. It is entirely capable of generating novel content that is far outside any individual precursor of its dataset, and that capability is what makes it a fantastic asset for creatives.
The thought I keep coming back to is that in order for AI to work you need people to create things. Would it be possible to get Andy Warhol without Andy Warhol ever existing to train on? My position is no.
That's kind of an interesting example, given that a lot of what got Warhol famous was the remixing and reapplication of commercial and popular imagery, as well as breaking from traditional concepts of what art was. Warhol was undoubtedly a rare creative talent, but he also worked extensively with base assets that others had created, finding new contexts for them that created a different meaning.
So, no, you can't get Warhol without Warhol, but you can be what he was and find a uniquely creative way to work in the novel media and contexts of the present.
There will be a constant incentive to create your own training data for these models to differentiate yourself from other artists. I'm confident there will be something of an industry where demand for new styles and whatever else spurs people to establish new art independent from the AI.
Or, long term, for AI to be creative enough we won't need a new Andy.
I'm a writer too and I've been seeing this sentiment a lot in the writing communities I'm in. My question is - why not use it?
I've always done all my book covers myself as I have a long history with visual creation and Photoshop. With the exception of one commissioned series I've always done all my own covers.
I've recently worked through my entire back catalogue and updated my covers with the help of AI, and I'm so happy with the results. My covers were always decent, but they've definitely gone to the next level now.
There are some that I've got away with just prompting, but by large this effort has still required a lot of Photoshop work on my part. Think of it as very involved digital collage.
I'm concerned because it's took a lot of work to get this done, and I think the results are great, but I'm getting a very real "Anti-AI" sentiment from a subsection of the writing community that isn't insignificant in size. AI is a tool for me, an extension for the creative process. Why not use it?
Humans create art. AI creates poor imitations, mostly trained on existing art that the AI creators didn’t have permission to use. It’s theft of the most insulting variety because it’s then being used solely to avoid paying or even talking to another human to create something beautiful. It’s killing true collaboration, it’s turning different artists against each other, and it’s tricking users into thinking they’re doing more than they are
It’s really depressing to see artists in different media get tricked into devaluing each others crafts like this. You ought to have more solidarity with other kinds of artists, especially if you want them to have solidarity with you
Humans are creating AI art. The creativity comes in from the person prompting the AI. A writer is not prompting an AI with ‘Book Cover’ they have a vision and are utilizing a tool to make it. Photography is art even though the camera has no creativity.
I disagree thoroughly. It's much closer to commissioning a piece from another person, because the only skill required to make it work is basic communication. You don't develop any visual art skills using an AI to generate art, it doesn't demand any artistic skill from you to produce something. You're not making the piece, you're asking the computer to do it
Would you consider photography as art, based on your set of criteria?
Edit: I see a longer comment on photography in your profile. You say that photography requires more skill, whereas AI is just kind of handed to you.
Have you tried using AI generation? I didn't see you mention it in the other comment, and your characterisation of how AI requires zero effort seems inaccurate to me, so if you haven't, I suggest you try it out. Mostly the things that come out are bad, irrelevant, or non-specific, unless you're pretty good at crafting prompts. It can take hours to make a good one. Even then, it takes a good eye to choose which of the hundreds of attempts to use as your final book cover or whatever.
I think this is pretty analogous to photography, in that most photographs are not noteworthy, and it's possible to take a photograph with zero effort, but skill, a good eye, and actually spending effort are the primary factors in what makes a good one.
I suppose it's just a question of what you want from a cover. I don't think AI covers are bad, necessarily, but you can't currently specify what I consider to be an appropriate amount of detail. My current project's cover features the protagonist, his face modeled on a particular 20th-century British poet, in a pose inspired by depictions of the bodhisattvas, eight hands spreading out with a particular significant object in each, with two of the hands posed in specific mudras - you get the idea. Maybe one day AI will be able to do that, but it is not this day.
You definitely couldn't get that kind of image from a simple prompt, but it would be possible with some strong base prompts, photo bashing and additional stock imagery. This is pretty much what I am doing with some of my more complicated covers right now.
Regarding using the AI-generated art for commercial purposes like a book cover, that may be against the terms of use of whatever site you're using to generate the art. It's also not copyrightable, whoch could potentially cause issues depending on your use case.
I would be interested in seeing what these look like, though I guess there's no way to do it without giving away who you are?
This is an interesting take especially when considering commissioned artwork. There are very real barriers to effective conceptual communication. I think in this circumstance it can also alleviate a certain level of collaborative frustration.
Some people are extremely effective at teasing out what they want, with alot of research and trial and error. I've predicted we'll be creating new jobs like 'Prompt master" for people who are especially talented at collaboration with these AI systems. They are incredibly nuanced and increasingly feature rich.
I agree with this, and I think we are already seeing some of the effects of it. Do you think creatives being effected should embrace and utilize AI in their work?
I think it's obligatory, to be honest. Part of why I'm so sanguine about the presence of AI in our lives is that it's a fait accompli at this point. The datasets exist, the models are trained and open-source, they're already going exponential improvement, and they're so damn effective that there's no force on earth that's going to put this genie back in the bottle. So, yes - creatives have to embrace it, or at least accept that it's going to be there alongside them and understand what that means for their particular industry going forward.
It's going to be an unpleasant shock for many, but I've said a few places that we're going to have a similar pre/post-AI divide in our lives just like we had a pre/post-Internet one. Think of all the things that changed because of the internet, and how the shape of our lives is different - and think of all the people who refused to engage with the internet because they disliked those changes, and how they fared.
I remember working a hotel front desk in 2006 or so and a man came in only to find that our last room had been booked by online reservation before he got in. He immediately flipped out at me and went on a 5m rant about how much he hated the internet, and you could tell that this wasn't the first time that his refusal to engage with it had robbed him of what he wanted. Now people like him are either dead, or they've reluctantly accepted using it because it's essential to life and they're getting snowed by scams and social media disinfo because they never developed the skills necessary to use the internet properly.
Something to think about, when contemplating AI. The manifestation will be different, but it's absolutely going to play out in the same way. People using AI assistants and AI generative tools will be more productive and desirable in most respects, and people need to engage - if only to the extent that they can recognize and defend themselves against the slew of AI-generated personalized content that will inevitably target them.
"Hey, were you at Target last weekend? You drive a grey CRV? You dropped a card from your purse in the parking lot..." Generative scam emails are going to get a lot more efficient in the future when someone can throw our entire super-comprehensive advertising profile at GPT-10 and watch it go nuts ferreting out our hopes and dreams so it can get its foot in the door. People need to have a basic understanding of the tech or they're going to get killed out there.
I agree. Even if certain AI models are discontinued because of the way their data set was trained, or when new laws are inevitably enacted to deal with copyright and authorship, the genie is out of the bottle and creatives will need to accept it's existence, utilize it as a tool and a part of their work flow, or deal with the ultimate consequences of keeping from their workflow.
When large companies like adobe begin adopting AI like firefly for their products, it becomes obvious the technology is becoming inescapable.
This isn't to say there wont be a place for traditionally created art. In fact, there may be a resurgence in the value of more traditional artistic creations in the face of readily available AI generated art.
So you just don’t want to have to talk to another human then? That’s why generative ai is good for you?
Reference the OP's reply above yours - conceptual communication is a hard task. Removing one step of translation from my understanding to an artist's understanding, and then again from their understanding to the page shortens the game of telephone. It's not about avoidance of a social situation, it's about making the path from my brain to the end product more direct.
Collaboration with other humans can be hard, true, but I think it’s worth doing
It depends on the task! In my particular case I want artwork about once every two weeks, which is a fairly rigorous workload for a commissioned artist. Not full-time, but definitely a significant chunk of their output. Setting aside that the cost of commissioned art on that frequency would be prohibitive enough to stop me from pursuing it, I would have to find a person who 1) draws in the style I want, 2) is open for commissions at a reasonable rate, 3) is prompt and punctual and 4) is willing to dedicate a large segment, if not the majority, of their drawing time to my project.
Now, money solves a lot of those issues but it is not trivial to satisfy points one through three with one person, to say nothing of four - and I can compromise on four, as long as I can find multiple people who fulfill the first three criteria, which is even harder to coordinate!
Involving other people in a project adds exponential complexity, due to scheduling, communications, etc. That's always been true, it's true in business just like in personal projects. Enabling one person to do more work makes that work more efficient, full stop. There is a lovely feeling to collaborative projects where everything lines up, I'll grant you, but in my experience those tend to be unicorns. The reality of collaborative work is that collaboration takes outsize time and effort. With AI, I can get the artwork I need done in the space of an afternoon, by myself, reliably, with no possibility of external factors outside my control disrupting my schedule.
as a writer I would choose to collaborate with a human over a computer. even if they have slightly different ideas than I do, that doesn't make them not valuable, and the intersection of our two imaginations can create something far more interesting than what I can do alone
as a visual artist, it seems like you see visual artists as an obstacle between you and the finished piece that you desire and I can't imagine that is totally separate from the rest of your problems
as an avid reader, if you don't have the linguistic skill to adequately describe a scene to a human artist, then I must assume that you also don't have the skill to adequately describe that scene to me, the reader, so maybe whatever you've written isn't worth my time to read
I'm happy that you have a clear ideal of what you envision the creative process to be! The reality of collaborative experience departs somewhat drastically from how you're describing it, in my experience. Human considerations of schedule, context and desire are seldom as amenable as we'd wish them to be.
You're free to hold whatever thoughts you want to hold concerning AI, and even to continue interpreting my replies in the least flattering possible light. However, increasingly I imagine you will be faced with people who find a valuable niche for AI in their creative process, their functional work or their lives in general. It may be prudent to try and understand their reasoning as they're presenting it, rather than interpreting everything in such a way that no change is required to the opinion you've already formed. Such a stance may lead people to conclude that it's not worth their time to continue a conversation with you.
It's going to be used to undermine the market value of creative skills, so I hate it. During a work presentation on AI, we were told that now "mediocre is free," which was supposed to make us feel better as creatives because our better-than-mediocre work would still have value.
But what I heard was, "the step below you is free, so how much more than nothing are you worth?"
It's also basically committing theft of IP on a massive scale, as companies train off all sorts of resources that include tons of copyrighted work.
I don't care how good it gets, if it ever becomes truly creative it will be creative in a way fundamentally different from humans, or else it will still simply be remixing our work into pablum. And it shouldn't replace artists, what a stupid thing to replace! I get the C-suite hates creatives more than anything because they can't fit the value of creative work into spreadsheets, and they see these as soft skills, so creative fields are an easy target if you want funding. But why not automate the drudgery out of work and let people express themselves with less strain on their time?
I hate it so much.
I have serious problems with scraping the web for these things - it's not unlikely anything posted anywhere ends up in a dataset somewhere, being used to train an AI for profit. There's probably someone scraping Tildes to use our comments to train an AI. I'm not a huge fan of current copyright law as it is, but I do think that if someone is going to use your work for profit, you should be compensated. And none of the people whose work is in Stable Diffusion or GitHub copilot currently are. These products that are being sold for a profit by entities that are using them to make a profit are built off the backs of people who spent a lot of time creating, who are now they're being told they are irrelevant.
I also think, with the current state of the internet and how training datasets for these things are created, that model collapse is probably inevitable for a lot of these tools in the long run. Unless the people curating data sets pay extremely close attention to filtering out AI generated content, eventually the models will succumb. And that will make anything they generate much less interesting than the existing output - which is already less interesting than the potential for human output.
This is also not a creative thing, but I see a lot of people trusting AI when they really shouldn't? I have a less experienced coworker who, instead of asking one of their more experienced colleagues, asking to see the reference books on my desk, or utilizing one of the many digital resources he's provided at my job, puts his questions about networking into chatGPT. And I'm like, okay, but AI will tell you completely wrong information and present it as fact - how do you know if it's not giving you bullshit? He didn't really have an answer for that. I saw this clip of some college students using AI to generate a final project for a class too, and it just looks like garbage. Like you are paying to be here to learn to acquire a skill, and this is what you decided to do? I have concerns about the proliferation of people effectively using it to shortcut having to learn how to do something, without knowing well enough to know if what it puts out is decent.
I've had some interesting conversations recently as an actor, because it seems like with AI, some companies are starting up the 'gathering data' stage. There have been a couple of casting calls put out recently for actors to come in for a few hours, be recorded, and their images/voices/etc. to be used for the purposes of AI.
The responses tend to be either that it's bad and nobody should do it, or that it should be heavily compensated - even moreso than in-perpetuity clauses. In-perpetuity clauses basically mean that there's no limit to how long and where an advertisement should be shown - most legitimate ads have for example 2 year buyouts, where they can show the ad for 2 years, but if they want to run it again they need to a) receive the actors' permission, and b) pay them more money. There's a general agreement that no actor should agree to in-perpetuity clauses, or at the very least that they should be paid an absolute ton of money - like in the hundreds of thousands of dollars. The growing consensus on filming for AI purposes seems to be going along similar lines to in-perpetuity clauses, which makes a lot of sense.
We've already had an instance of a lawyer using ChatGPT to generate documents he submitted to the court ...it went exactly how you'd expect. ChatGPT is impressive but way too many people think it's a source of information and that's a HORRIBLE way to use it.
Here's the thing; those tools all did have massive effects on all of those industries.
When is the last time you paid someone to make a handwritten copy of something?
How many engineers got fired because one person with a CAD program made them redundant?
VFX has also been ruined by digital effects, though indirectly; now that they are all interchangeable and no longer linked to the rest of production, a lot of work just goes to the lowest bidder. Though admittedly there are many other factors at play.
Traditional animation was absolutely trashed by CGI; literally the only Acadamy Award for animated film to go to a 2D film is 2002's Spirited Away, which could not have been produced in America because the industry in this country wasn't big enough for it even then. Disney is a giant international multimedia conglomerate and they have long ago shuttered their 2D animation studio - twice!
I don't think you're wrong, exactly, but the thing that these disruptions tend to do is make those jobs available to fewer people who may or may not be somewhat higher paid for a brief moment, and that is certainly a negative effect.
I think it should be noted that disruptive technology in the creative world demolishes the low and middle tiers of the market, leaving the high end.
Canva and other DIY creative tools have enabled a lot of small business owners to make their own logos and graphics. But big companies and orgs still want professional designers and super polished outcomes.
The proliferation of consumer camera technology has destroyed small photography studios. Most people do their own family portraits and professional headshots. But people still want professional photographers for important events like weddings.
I think that generative AI art will be a huge boon to enterprising artists who work in the high end of the market, like making conceptual art for movie and game studios, for example. It'll enable them to achieve ever higher heights and greater volume.
But artists who aren't at the top are going to get wiped.
I'm not saying I'm not learning this crap, I have to. I'm also not afraid, I'm pissed. I'm pissed that there's more fuel added to the fire that creative skills can be had cheaper because they're soft. I'm pissed at every exec and tech-bro scammer who's out there saying content is free to create now. In part because of what it says about their opinion of creativity, and in part because it reveals them as either grifters or scammers. This tech has been massively expensive to develop, and right now that cost is being borne by VC money. At some point, it's going to stop being cheap. At some point, once they've displaced enough creatives that it's more expensive to rebuild human creative teams than to pay the fees, AI is going to cost as much as or more than human labor.
And I'll also say this. Right now, all AI is capable of doing is chewing up and regurgitating the contents of its model. It can't be more creative than humans, it can't add creativity. All it can add is more average, because that's what it is: an average across a bunch of existing work.
I'm learning it, and part of why I hate it is because I'm deep in it, and I see how bad it is at so much of what it promises to do.
Instead of a bad doodle on a napkin, now it's an intern typing into chatgpt. A program that just wants it to like you, no matter what. It needs those tickets.
I don't know if I shared this with Tildes before, but I came to a conclusion about AI art a while back.
I'm not worried about it.
What makes any given thing "art"? What is the fundamental difference between the layered strokes of paint that make up a white-colored wall and the layered strokes that make up a painting? Intentions. The transmission of thoughts and ideas.
AI artists not only do not have those things, they literally cannot even start working without a human being providing them first. And even then they are limited to what we can describe with words - which is a small subset of the thoughts and feelings we can have - and restrained even further by the AI's understanding of those words.
Look at
these
three
images.
Guess what they are supposed to represent. Here's a hint: they're all supposed to represent the same thing.
Did you guess "A picasso painting depicting the emotional scarring a woman feels as a result of a miscarrage"? Because that was the prompt I gave Stable Diffusion.
On the other hand, take a look at @kfwyre's commissioned artwork he shared with us not too long ago. He spent money on having a real human make that, and you don't have to read his story to tell what the intention and feelings behind it are.
Aesthetics are also an important part of art, for sure, but right now that's all that AI can produce. Until we can get that extra level, AI art is nothing more than a useful illustrative tool.
When I look at the third AI image, I get a deep sense of grief. Obviously I have no idea what that grief is about, but I think that's ok.
I think AI is just another tool that lowers the skill floor for people to generate art. I don't think it's any different than the software that the artist used to generate that commissioned artwork.
Sure there are questions about ethics with regards to scraping, but I don't think scraping itself is unethical. I equate it with a person going to a museum and seeing different pieces with different styles. If I commissioned an artist to paint something in Picasso's style, would they be committing fraud? Or if I bought a replica of the Mona Lisa, is that fraud? I think the ethical questions come into play when people misuse AI to commit fraud, pretending that something is human made and selling it as such, or selling someone's art style as their own. But that's not the intended use of AI, so I can't put the blame on it. If I painted an exact replica of the Mona Lisa and sold it as an (or the) original, would that make paintbrushes unethical? Or museums?
I have not used AI in any creative process. I make my own artwork and, when appropriate, leather and wooden goods. I believe the role of AI simplifies and automates the creative process. It neither enhances creativity nor limits originality. No, it essentially provides a cheaper alternative to commissioning artwork from an artist.
Fact of the matter is, AI artwork looks remarkably decent and it's going to put digital artists in a bind if that's their only source of income. There is the concern that some might use AI programs to feign true artistry, as if they painted the image themselves. That is perhaps one ethical dilemma in the future of AI artwork.
Consider this moment to be the mass production of yet another industry that required careful and experienced hands, like how the plastic jug replaced glass bottles, or how plastic shoes put cordwainers and cobblers out of work. Digital artwork has its days numbered.
What I think will save art (until moving, blinking, talking robots can copy a Mona Lisa stroke by stroke) will be a resurgence and increased value in traditional art. To be able to hold the thing that someone made by hand off of the screen of your laptop or smart device, to see it with your own eyes under different degrees of light, that will be the culture that needs to develop against the one we have now that values convenience and price. I don't think it will be easy to get there, but there should be a push for it.
The analogy I've used in other conversations is typesetting. Back in the day it used to be a specialist industry, with large equipment and years of skill needed to do it properly. Typewriters changed the industry dramatically, but MS Word and professional programs like InDesign were the death knell for pretty much all non-digital typesetting.
Yet for things that matter, you can still find people who will do traditional typesetting. It's a boutique industry now, but it exists and people make their living in it. Traditional typesetting is a prestige service, while most use cases for it don't really care how the text is generated so long as it's legible.
I mean, I think even if a robot could copy the Mona Lisa stroke by stroke it still wouldn't be art. It wouldn't be the original painting and it would lack that painting's history and context. As Walter Benjamin argues:
Benjamin talks a lot about the 'aura' of a work of art, which he says a production necessarily lacks because of this lack of context. But it's also more than that imo - at least with regards to AI. An exact copy of the Mona Lisa might provoke emotions in people who view it but that doesn't make its creation art. The robot hasn't gone through a creative process to make the copy, they haven't been moved to convey an emotion or a scene to other people. A reproduction isn't art because it's in no way new or vital. The reason we revere the Mona Lisa isn't because of the brushstrokes and pigments used by Da Vinci, it's because it was made by a person at a particular moment in time who felt compelled to paint this one particular woman in such a striking way.
Copying the Mona Lisa, or asking AI to create images 'in the style of' certain artists is simply a pastiche, the ultimate style over substance. I hope you're right about traditional art being more valued as a result of the increase in AI art but I'm worried it's going to be an uphill battle of restating art's inherent value over and over again to people who just want something which fits a certain aesthetic and don't really care about its provenance.
The statues, monoliths, sunken ships and their treasures that we unearth to this very day are a testament to the value of traditional art. It exists because of us, and yet, it is independent of us because it transcends our limitations, our values, restatements and thoughts, and survives us. Naturally, it doesn't matter what we think of it because it will last longer than us. It matters even more if it is destroyed, for it was a message deemed somehow threatening, despite its antiquity, to the destroyer. If it is stolen, it gains modern notoriety and appreciation.
Traditional art will always be valuable.
If anything, the emerging vinyl market is a good example of your thesis.
It honestly depends. I am a digital artist, I make 3D sculpts, drawings, and pixel art. The past week I have messed around with stable diffusion to get specific poses generated for potential sculpts - and it has been a great success. To quickly be able to visualize what I specifically attempt to emulate has been a massive time saver.
To add onto that: The hostile Pinterest flooding of Google Images has had a horrible effect on finding specific search queries - not to mention the quality of web search these days in itself.
Using the images generated for reference is in my opinion valid. It's in essence the same as using a picturebook, or perhaps your favorite screenshots from a cirque du soleil performance, to create something new.
Prompting does take some effort to get exactly right, and all 'good' AI 'art' is heavily edited in post, which does take knowledge and skill. Appeal remains subjective. See it more as collage art than traditional.
Where I draw the line is the masses who like to go 'I press a button, now I'm an artist'. Putting an image into a filter takes no skill, and is little better than messing around with the liquefy filter in Photoshop.
There are plenty of people who love taking shortcuts: anyone can emulate Pollock's or Mondriaan's techniques, but most of them won't be able to tell you why they have value and their copies do not.
Sadly, corporate does not care. I would advice any artist or writer to have a look at AI and get an understanding of its pros and cons, and make it a part of their workflow rather than be staunchly opposed. Stay ahead of the curve.
Oh boy. I was an SEO writer for a company that built websites for local businesses. Their CEO, whom everyone treated like a cult leader for some reason, tried generative AI last year around Christmas and that was it. He loved it more than he ever loved any technology ever. He made me start using it to write all web content including company blog posts. I pushed back. He made me put my freakin' byline on it. I pushed back harder. I got fired. (Fun fact: he's also trying to expand his local "Jesus-Football-Lots-of-White-People" print magazine into other markets, so he figured he'd save on costs by replacing me with AI. Not that I'm petty, but I hope they start hemorrhaging cash with their stupid print magazine.) The bottom line for me is: generative AI is fine if you don't normally write and need to compose an important email or whatever. But as it is now, it generates hackneyed crap instead of good web content.
Do you see any professional value in it at all? Would you use AI as a tool to help your own writing? Tasks like generating ideas, rewriting a thought, looking for counter points, or maybe to assist with a general outline for a large piece you were writing? or are you of the opinion that it has no space in professional writing or content authorship?
I speak for myself only.
No.
Certainly not.
No.
No.
I think it were a worse world and a worse society which admitted the use of these things. I have no doubt that you can find metrics which are improved—perhaps even monotonically so—when people use them.
I think I understand what you are saying here. I appreciate your honestly and your defense of the human creative process.
Ultimately its just a new tool and I imagine the fears of it aren't dissimilar to when photography first became widely available. Ultimately it won't replace anyone, at best it will open art to new directions and at worst make art more accessible to people who weren't or couldn't ever spend money on it anyway.
As a creative, I have mixed feelings about the impact of AI on the future of art, illustration, and design. I approach technological innovations with weariness, fearing that they might replace human creativity and diminish the uniqueness of art. I also recognize the potential of AI as a powerful tool in our creative processes. I choose to embrace innovations and incorporate them where they enrich my work, such as generating ideas and streamlining workflows. It has been essential to maintain a balance and ensure that AI remains a complement to creativity rather than a replacement.
I m new to Tildes and had commented in the board game group about using AI to create a board game from scratch. The journey has surfaced how much humans still need to be involved in the process.
Have you used any AI tools in your work flow? There seems to be a new tool with a unique application popping up daily. https://theresanaiforthat.com/
I have... mixed feelings. On one hand, as an amateur digital artist, I definitely don't appreciate the thought of having an AI model trained off of something I put hours of work into to spit out a facsimile in a matter of minutes (though I'd be very surprised if my art was considered good enough for someone to feed it to one of those models, lol), and it bothers me on a philosophical level that I can't quite pin down concretely - perhaps irrationally, yes, but it still just doesn't sit well with me...
...but on the other hand - I think of myself as being much more skilled with words than I am with visual mediums, and I can immediately see some large benefits to the capabilities of AI in that regard, as a supplement to the creative process. Being able to use an AI model to quickly create "concept art", if you will - to get a quick feel for a scene's mood, for how a setting might look, for a character's appearance - is potentially a significant boon in my eyes, particularly for independent creators.
I'd like to emphasize the concept part of that, though. Right now, I can't see myself ever feeling comfortable directly using AI-generated artwork as part of a user-facing project. As a personal resource to help myself conceptualize and visualize a scene or setting or some other imagery in my head? Definitely. As e.g. the cover artwork for a novel, or a background in a game, or anything like that? Definitely not. I would say @Akir nailed it in describing AI art as a useful illustrative tool. Those are the benefits I see in the concept, and generally speaking, that's something I'm okay with.
I hate it. It’s training people that art can be cheap and easy and it’s not even worth talking to another human for. Theft of art is like nothing because art is like nothing. The time and effort it takes to get good is utterly devalued. Dark days indeed when people can’t even work together on an art project anymore
Generative AI can’t replace humans for creativity and it’s disrespectful to ourselves that we’re even trying. It’s as if humanity is losing all confidence in itself
I appreciate your passion and defense of the craft. Do you think this AI paradigm shift is inherently different to past creative / art paradigm shifts? For example, The introduction of photography or the more recent advent of photoshop? If so, why?
Would you ever adopt AI in your creative process on any level? I recently started using Gigapixel AI to upscale images for production purposes. Are you also against this type of AI application in the creative process?
Edit: A word
I do think it’s different.
A camera is faster than painting, yes, but there’s still a lot of work that goes into photography. Even now with the majority of people carrying around cameras at all times, good photography takes practice and work to get good at. There is still much to learn about composition, lighting, angles, the mechanics of the tool itself.
Photoshop makes some things possible that are not so easy in physical media. However it is still a tool that requires knowledge and practice to use effectively. You have to have not just the idea, but the skill to turn your idea into something visual.
Take it a step further even and let’s talk about carpentry. There are insanely fast, even robotic tools like CNC saws and mills, that can take an awful lot of the physical toll out of crafting furniture from wood. However you still have to understand how to design and build furniture in order to use those tools to make anything worthwhile
Generative AI requires nothing from you. It uses nothing OF you. It takes no time, no training, minimal effort, it deludes you into thinking you made something when actually it was generated by a computer from the fruits of others efforts.
When you create something of yourself, especially artistically, there should be some kind of sign that you created it. Everyone has a personal flourish hidden within but generative ai can only derive from what it has seen, so you end up with a cheap copy of someone else’s flourish. Whether visual or written.
to bring it back around that’s effectively what I see as missing from many current applications of generative ai: other tools facilitate you in creating, but they don’t take it out of your hands, so that piece of you that the art came from, still comes through in the final piece. Prompt based generative ai takes the entire process after the prompt out of your hands - it won’t have your flourish, or any intrinsic sign that you made it - because you didn’t
Probably not. I prefer to do the things myself even if I am not completely good at it, because that’s the only way to learn and grow as a human
It is interesting that you bring this up because I am not wholesale against AI-type tools. In this case, the algorithm hasn’t taken over any of the creative process from you, it is simply making your image larger. If I were against this I would have to be against using bicubic sharper when resizing images too, and I’m not.
So I guess that’s where I draw the line: helper tools that make things easier, but don’t ultimately remove the artist from the art, these are proper tools. But if the “tool” seeks to replace the entire visual artist: that I have big problems with
I’m kind of fine with it.
I’m a writer and I mostly use it for visualization purposes. Since I primarily write fantasy, it’s nice to be able to create the scenes in my head in an image to reference in some sense. It’s never perfect and many times it’s more work to explain to it what I want, but it’s cool.
I largely view AI as a tool to get the less interesting parts of the creative process out of the way. I often think to popular manga artists in Japan who release weekly chapters, and sometimes kill themselves trying to deliver their work on a weekly cadence. If they can use AI to develop their backgrounds and take out some of the more time consuming parts of the process by all means go ahead. It gives them the ability to focus on their story, its characters, the framing, etc.
I think AI is only really going to allow the creative themselves to take more time on the intention behind the art or work they are creating. I think a lot of people discount the intention of art.
There are definitely ethical implications to the use of AI, but I don’t think rejecting the technology is the way to go about making sure things remain “above board”. The tech is here and it isn’t going away. Best to stay involved and part of the conversation than to be left behind and let others dictate how the tech affects my work.
How do you feel about AI being able to write novels?
I still don’t feel one way or the other. Much in the way I don’t find AI art all that inspiring, it’s the same for written art. I think intention is the most important element when we are talking about art.
AI doesn’t have experiences, memories, thoughts, or anything like that. All creatives draw from their collective experiences, memories, feelings, etc. anything. I think an AI could probably create a whole novel, but there is just more to art in general than being able to organize a bunch of random thoughts together in a readable manner.
Do I think it will always be this way? Maybe not. I follow the discussion around AI pretty closely, so only time will tell really.
I'm off two minds about it, but the genie is out of the bottle when it comes to this technology becoming more powerful and intuitive. I can't wait until a feature length film comes out from a 13 year old with a render PC and an iPhone that makes a splash at film festivals, but at the same time, procedurally generated Spider-Man films are probably in our future that will put a whole lot of people out of work.
It's definitely a paradigm shift, and we're probably just scratching ths surface in terms of its evolution and future applications. Midjourney has spoken about their plans for generating 3D models from text, and Adobe also has a bunch of interesting AI powered features like photoshops generative fill, which was talked about in another post or Firefly's unreleased features like text to vector, text to brush, and sketch to image.
Personally, I won't use AI for any creative tasks where I want the benefit of the practice for myself and my own cognition. I want to train 'Me' to artistic mastery, not an AI. For semi-creative tasks in my 'day job' as a content manager, where I don't care about achieving mastery, I would be OK with using AI tools.
Context: From the perspective of a writer and game designer. I have commissioned art and covers before, and due to my economic station - will continue to do so.
First, I think it's important to do away with concepts of like/dislike, and I worry that harping on ethics/theft is growing to be denial of reality. At this point, the technology is developing and proliferating faster than any reasonable organizing body will be able to properly adjudicate. In addition, I think it's simply idealism to believe that soulless corporate entities will ever go back to a "pre-AI-Generation" world. These tools will become more advanced, and the corporations with the money will ensure that AI-training and modeling won't be regarded as IP theft.
So, if I accept that the giga-rich will be using these tools to maximize their bottom lines, and if I choose to accept that fact as incontrovertible (you may choose otherwise). I can no longer afford the ethical dilemma of "should this be used". Doing so would be simply giving a powerful advantage to the rich and powerful, and denying it to the poor who are struggling to compete. People online would have me boycott a local butcher for using AI art to make his ads, because of the societal value of art. I posit, that this is a tax of economic status and privilege. Wal Mart and Amazon will not crumble under an "AI Boycott", so the only people who will suffer are the people who we should want to survive.
I understand the distress artists feel to have a competing technology entering their field - much as I understand the distress accountants felt when you no longer needed to pay them to do your taxes.
The value of art is getting discussed a lot lately in this conversation - but I think many people are dropping the nuance of "Value =/= Cost". Artists are seeing something many industries have seen before. Technology is driving the "Costs of your specialization" down, and increasing competition - this is a natural and common phenomenon - but it's not one we can simply deny. It also doesn't mean that art and artists have less value.
I'm a toy designer for The LEGO Group. We have a digital building tool for design called LDD-Pro. It has all the LEGO elements we produce available. It has taken more than a decade to develop and every brick is supposed to have all 'legal' building possibilities and to not allow us to place them 'illegally' (in impossible positions).
It is great, but virtually everyday we find something is possible that should not be, and possible positions are not available because we have come up with a new way of using the bricks. If an AI tried to build with the essentially infinite possibilities available it would currently not work, too many things would be placed where it would not be possible to do so in reality and shapes would be made from weird choices of elements. But I guess it would learn. To a point.
It would then run into issues with stability - can the toy be played without falling apart? Can a mini-figure be put in a car without the floor being pushed out, can the wheels turn, can it be pushed without flying apart? Doors open on a house? Legs move on a Mech without popping parts off? Drill turn on a digging machine? I guess with time this too can be taught.
Then it runs into building flow. Is the model stable enough to not be frustrating during building? Is it satisfying to build? Is it fun to build? Now it's getting subjective, but I guess again this can be learned with enough data.
Then there is the totally subjective aspects. Does it look good? Is it appealing to children? Is it cool? (And to a lesser extent, but true for a fair percentage of our products - is the license holder (Disney/Warner/Microsoft/Hasbro/or whoever) happy?) If they are not, what are you going to change?
I think a lot of this makes this particular design job one of the last ones that AI will be able to do. I hit retirement age in 18 years, hopefully I won't be replaced by the machine overlords before then. :D
I’m a graphic designer and I’m more afraid of what AI will do to our heavily capitalistic societies than I am about its impact on my work. We should be worried about union busting, copyright infringement, where this data is stored, how data is being collected, who has access, privacy, internet scraping, and proprietary information being leaked. Just for starters.
I think AI will streamline some of my workflow. I won’t have to fuss with clone-stamping for more skyline when the photo I have to use isn’t big enough, I can just render more. While there are fears of devaluing my abilities I think what will happen is that AI will just be the new starting point for me to refine a client’s concept.
I also think that AI, like Canva and Adobe Express, will begin to generate a lot of homogeneous designs. People will pay for originality and interpretation, and that has always been what creativity has been about. Google slides did not turn people into presenting geniuses, Photoshop did not make everyone Ansel Adams, Canva is not going to make everyone a layout expert and AI is not going to render creativity moot.
Definitely an insightful perspective. I'm sure AI's impact will be felt throughout society as we move into the future. Sometimes for good and sometimes... maybe not so much.
I will say that in terms of copyright infringement (which is a very important point) these large learning models already exist as a proof of concept regardless of the data they are trained on. Pick your favorite copyright infringer and shut their business down. It's not going to stop the technology or new models that are trained ethically. Adobe's Firefly is a good example.
I agree, and I can attest to the positive changes to daily work flow. I try to look at AI as a tool to enhance my workflow and remove redundancy.
Very insightful. I think this is an important point that needed to be made.
Yes, I have used Midjourney for my digital art hobby, and ChatGPT for both writing and programming. Midjourney is delightful, albeit problematic in its datasets. ChatGPT (3) is an okay way of getting a framework, sometimes. It's borderline terrible at creative writing, and barely competent at simple coding tasks. I'm only just learning C# and I'm better at programming sensibly than GPT3 is. I'd be interested to see if something like GitHub Copilot is better, but I'm not willing to pay for it.
For me, personally, it's a source of inspiration. I am functional at the act of creation but I often lack motivating ideas of what to create. The AI art I generate rarely gets seen by anyone but me - I use it to create harmonic color palettes, poses to utilize, or like I said, subject ideas. In a way, it feels bittersweet, because the AI images are far beyond my current talent level. It makes me sad that I don't feel comfortable sharing them due to the negative discourse surrounding anything "AI" and art.
For me, the former! It doesn't stop me from creating the original ideas for art pieces that I come up with. It's only when I'm uninspired already that I turn to it. And in honesty, it's not as if my original ideas are groundbreaking either. I mostly draw cute girls or animals - hardly original. I believe that the people with strongly original ideas will be driven to create those ideas regardless, and can't be limited by AI in that sense. But maybe overall there will be less strongly original art, because those of us uninspired artists might rely on similar ideas to create art of?
This is the big question, of course. I believe that it depends on how the AI is trained and how the art is used. If the AI is trained on art from folks who didn't consent to having their art used, if the prompt deliberately targets one of those artists, and if that result is used for profit without significant effort from the prompter, I believe that's unethical. But if the artists consent, if the prompt isn't targeted at a specific style, or if the result is used as inspiration rather than used as-is, that blurs it quite a bit. I am in the group of people that believe a style cannot be copyrighted, and that apeing styles is an important learning tool. But it sits poorly with me that you can tell Midjourney to make you a Loish-style print and then undercut her by selling it on Redbubble without her consent or permission, or giving her credit/revenue share.
I want AI to be a useful tool for artists. I want it to be an aid for people with aphantasia to picture things. I want folks with hand pain or tremors to be able to create art with it. I think it has the potential to be a powerful leveler of the playing field, but only if capitalistic greed doesn't ruin it first.
The biggest proponents of AI art and machine generated text are very clear that they intend to working creatives out of a job. It's driven by an incredibly petty and childish jealousy of people who have put in the work to learn the skills by those who haven't. Ned Ludd was right.
Art sold for display, as long as you know what is and isn't AI generated, I see no issue since human made art will likely be valued far higher than AI art which will become rather common.
The rest who art for a living will probably have to embrace AI. Use the AI to create about what you want and then use your skills to tweak it to your liking to make something better than those who just use the AI. Depending on what it is, it might save hours of tedious work and it'll be up to you what part you'd apply a personal touch to.
I'm curious if new copyright laws might be added to apply to AI only. For instance, it'll be too easy to copy another artist. Imagine someone uses AI to copy your unique art style so instead of hiring you for custom work, they just inject the AI with copies of your art and make it without you.
The good things I see is better art for things that often lack it due to cost issues like pictures in books or models in video games (more unique models rather than a bunch of re-textures :D)
I used the prompt "What would be a good 20 year forecast of the impact of artificial intelligence on the future of art, illustration and design"
CHAT.OPENAI answered
Predicting the precise impact of artificial intelligence (AI) on the future of art, illustration, and design over a 20-year period is challenging, as technological advancements and their consequences can be unpredictable. However, based on current trends and developments, here are some potential areas of impact:
Enhanced creativity and collaboration: AI tools and algorithms can assist artists, illustrators, and designers by providing new tools, inspiration, and possibilities for creative expression. AI can generate ideas, suggest compositions, and even generate entire artworks based on learned patterns and styles. Collaboration between human artists and AI systems can lead to new and unique artistic outcomes.
Automation of routine tasks: AI can automate repetitive and time-consuming tasks in the creative process, such as colorization, image retouching, or layout design. This automation can free up artists' time and allow them to focus more on the conceptual and artistic aspects of their work.
Augmented reality (AR) and virtual reality (VR): AI can play a significant role in enhancing AR and VR experiences. AI algorithms can analyze real-world environments, interpret user interactions, and generate interactive virtual elements or characters that seamlessly integrate into the user's surroundings. This opens up new possibilities for immersive artistic experiences and interactive design.
Personalized and adaptive artwork: AI algorithms can analyze vast amounts of data about individual preferences, artistic styles, and cultural trends. This information can be used to create personalized artworks and designs that cater to the unique tastes and interests of individuals.
Ethical and societal considerations: As AI becomes more involved in the creation of art and design, ethical questions will arise regarding authorship, ownership, and attribution. Discussions around the originality and authenticity of AI-generated works will become increasingly important.
Changing the nature of work: The integration of AI in the creative process may lead to shifts in the roles and skills required in the fields of art, illustration, and design. Artists and designers may need to adapt to new ways of working, collaborate with AI systems, and develop skills in understanding and working with AI tools.
Impact on traditional forms and new possibilities: AI can challenge traditional notions of art, illustration, and design by introducing new forms, styles, and techniques. AI-generated art can push the boundaries of creativity, introducing novel concepts, and sparking new trends.
It's important to note that these are speculative projections, and the actual impact of AI on the future of art, illustration, and design will depend on various factors, including technological advancements, societal acceptance, and the choices made by artists and designers themselves.
As with any forecast, it's important to remain open to unexpected developments and be adaptable to the changing landscape. The creative field is likely to continue evolving in exciting and unforeseen ways as AI technology progresses.