Oct. 23rd, 2024 at 9:14 AM
Have we become AI?
Some reflections on Artificial Intelligence, its impact on illustrators
...and trying to answer the question: What do we do now?
Summer 2024, Cover artist: Yuji Takahashi
By Sarah McIntyre, first published in The Author magazine,
issued by The Society of Authors, commissioned by editor James McConnachie
Reprinted with permission
Let’s talk about Artificial Intelligence. How do you know AI didn't write this article? An AI text generator could do a pretty good job of amalgamating words of enthusiasm and outrage about itself that it’s scraped off the Internet. You’d get a fairly convincing mash-up of what people are saying about AI right now.
But the outrage is real: translators, scriptwriters, poets, audiobook narrators, illustrators and writers fear AI will replace us. And from what I've heard from illustrators, what angers them most is that it's our own work competing against us, that these AI generators operate by 'scraping' the work we've posted online to train and develop its products, without crediting or paying us. Do read the latest SoA Resolution on Artificial Intelligence approved by 97% of voters at the last Extraordinary General Meeting. In it, the SoA statement refused permission of copyright-protected works to developers without specifically agreed licensing arrangements. It called for developers to provide full transparency on works used in development, to be specific about what they're requesting, to ask permission from the relevant rightsholder, to pay and credit authors, and to remove any works that developers have already fed into their systems.
The SoA's survey in January noted of its members that 26% of illustrators had already lost work due to generative AI, and 37% of illustrators said it had caused the value of their work to sink. The survey also revealed that 12% of illustrators have used generative AI in their work, 31% of illustrators have used it for brainstorming ideas, and 5% have used it 'because their publisher or commissioning organisation asked them to'.
While the SoA battles for credit and pay, illustrators are left thinking, what do we do now?
When I set up the Pictures Mean Business campaign* in 2015, I'd been listening to veteran illustrators saying that no one ever really listened to their complaints about lack of credit for their work. Decades of objection hadn't helped at all. So I thought, we have to stop banging our heads against a wall, we need to approach this differently to get results.
To start with, which objections to AI won't work? The argument that AI takes away our jobs may be about as effective as it was for dock workers in the 1960s. As illustrators, we've even benefitted from the internet forming databases out of other people's copyright-protected work without permission: who hasn't used Google Image Search for reference material? As Darrell Maclaine pointed out on X, "You've all been using that without comment for 25 years. I don't see the fervent calls to 'hire real photo librarians!'"
A second argument also doesn't fly: imploring the publishing industry and the general public to fight illustrators' corner out of love and respect for us. When hard decisions are made that involve money, business just doesn't work that way. Even on a small scale, a cash-strapped book festival or an overworked illustration agency may rely on AI images when they can't see any other way of making ends meet.
But, at least for this small scale of UK publishing, book festivals, schools and libraries, I think we can flip the argument. If we stop making ourselves the focus, and make it about other people, we have more chance of convincing them. With Pictures Mean Business, I tried to show people how everyone benefits when they credit illustrators properly for their work. People love to hear how they can get an advantage in publishing, how they can make their books more searchable, how they can tap into larger illustrator fan bases to get publicity for their books or festivals, how they can better inspire children to read.
How could this work in relation to AI?
Art by recognised illustrators make powerful branding, people are drawn to it.
I recall crossing a street in Edinburgh to look more closely at a poster featuring Judith Kerr's cat; it instantly brought up recollections of my favourite childhood Mog book and I even tweeted this photo of it. Now imagine you're a book festival organiser. You have the choice of coming up with a poster that is nice in a sort of generic way, or you can feature Chris Haughton's dog from Oh no, George! or Nadia Shireen's Billy and the Beast, or Melissa Castrillon's wonderful organic patterns. The first choice - the AI image - expresses disdain for illustrators (which they will note), and the second, besides tapping into a brand, lets you show off your connection with an amazing creator and highlight how your team support artists. If that's the mission of your festival, this absolutely makes sense: it makes you look good, draws extra attention to your festival, and taps into the artist's own fan base.
For writers, associating with real artists boosts your reputation. Illustrators scrutinise authors before they agree to take on their books. If you have credited artwork on your website (instead of uncredited, or AI), a top artist may be much more inclined to work with you, and to big up your project afterward.
But AI issue poses deeper questions about what it means to be a creative person. What can we make that a computer can't?
Over the past decade, I've seen social media turning so many of us into something that might as well be generated by AI. Illustrators copy online illustrators who have copied other online illustrators.
Our artwork isn't only about creating a certain style, it reflects ideas, and communicates how we see the world. And I'm not sure how much we really ARE seeing the world, and how much we're looking at it through the lens of other people, because that's what's acceptable in our bubble or peer group. It's easier to accept what people whom we admire say about an issue than to dig deep, seek uncomfortable truths and details, and draw something that freshly interprets what we've discovered. Social media works like ChatGPT; it aggregates us in a way that suppresses anomalies and outliers, and comes up with a convincing-sounding average blend. But that uniform result, just like ChatGPT, gets things wrong, and sounds boring. Behavioural science expert Rory Sutherland argued that if, using AI, 'you could produce someone who is the average of all your friends, you probably wouldn't like them very much and they'd be extremely dull'.
'But we just need to make a living', you could reasonably argue. 'If publishers are playing it safe, and I'm turning out the bland content they want, at least it pays the bills.' But this is unsustainable, what distinguishes us from AI? And why should anyone care if we disappear?
I don't think there's one solution to this problem, and perhaps therein lies its beauty. You need to find your own way, discover what makes you and your work unique.
Perhaps this involves stepping back from the computer for decent lengths of time, building off-line relationships with people, and learning how to make images that have nothing to do with anything going on in your online peer group. Try your hand at something that has nothing to do with a suggestion from an internet influencer, and don't worry about how it will look on social media. Explore it and revel in its unique quirks and eccentricities. For me, I've recently enjoyed drawing scenes of local community life; people I've been getting to know - they love seeing themselves in pictures, and I can draw them in very specific locations, with an awareness of their personalities and loveable quirks that a computer hasn't yet cracked.
Writer and illustrator Sarah McIntyre is currently creating the Adventuremice series with her co-author Philip Reeve and is trying to spend less time on the internet. And yet, there she is: jabberworks.co.uk
*Pictures Mean Business website generously created and funded by freelance artist Soni Speight. #PicturesMeanBusiness hashtag created during a discussion on Twitter with artist James Mayhew.
Summer 2024, Cover artist: Yuji Takahashi
By Sarah McIntyre, first published in The Author magazine,
issued by The Society of Authors, commissioned by editor James McConnachie
Reprinted with permission
Let’s talk about Artificial Intelligence. How do you know AI didn't write this article? An AI text generator could do a pretty good job of amalgamating words of enthusiasm and outrage about itself that it’s scraped off the Internet. You’d get a fairly convincing mash-up of what people are saying about AI right now.
But the outrage is real: translators, scriptwriters, poets, audiobook narrators, illustrators and writers fear AI will replace us. And from what I've heard from illustrators, what angers them most is that it's our own work competing against us, that these AI generators operate by 'scraping' the work we've posted online to train and develop its products, without crediting or paying us. Do read the latest SoA Resolution on Artificial Intelligence approved by 97% of voters at the last Extraordinary General Meeting. In it, the SoA statement refused permission of copyright-protected works to developers without specifically agreed licensing arrangements. It called for developers to provide full transparency on works used in development, to be specific about what they're requesting, to ask permission from the relevant rightsholder, to pay and credit authors, and to remove any works that developers have already fed into their systems.
The SoA's survey in January noted of its members that 26% of illustrators had already lost work due to generative AI, and 37% of illustrators said it had caused the value of their work to sink. The survey also revealed that 12% of illustrators have used generative AI in their work, 31% of illustrators have used it for brainstorming ideas, and 5% have used it 'because their publisher or commissioning organisation asked them to'.
While the SoA battles for credit and pay, illustrators are left thinking, what do we do now?
When I set up the Pictures Mean Business campaign* in 2015, I'd been listening to veteran illustrators saying that no one ever really listened to their complaints about lack of credit for their work. Decades of objection hadn't helped at all. So I thought, we have to stop banging our heads against a wall, we need to approach this differently to get results.
To start with, which objections to AI won't work? The argument that AI takes away our jobs may be about as effective as it was for dock workers in the 1960s. As illustrators, we've even benefitted from the internet forming databases out of other people's copyright-protected work without permission: who hasn't used Google Image Search for reference material? As Darrell Maclaine pointed out on X, "You've all been using that without comment for 25 years. I don't see the fervent calls to 'hire real photo librarians!'"
A second argument also doesn't fly: imploring the publishing industry and the general public to fight illustrators' corner out of love and respect for us. When hard decisions are made that involve money, business just doesn't work that way. Even on a small scale, a cash-strapped book festival or an overworked illustration agency may rely on AI images when they can't see any other way of making ends meet.
But, at least for this small scale of UK publishing, book festivals, schools and libraries, I think we can flip the argument. If we stop making ourselves the focus, and make it about other people, we have more chance of convincing them. With Pictures Mean Business, I tried to show people how everyone benefits when they credit illustrators properly for their work. People love to hear how they can get an advantage in publishing, how they can make their books more searchable, how they can tap into larger illustrator fan bases to get publicity for their books or festivals, how they can better inspire children to read.
How could this work in relation to AI?
Art by recognised illustrators make powerful branding, people are drawn to it.
I recall crossing a street in Edinburgh to look more closely at a poster featuring Judith Kerr's cat; it instantly brought up recollections of my favourite childhood Mog book and I even tweeted this photo of it. Now imagine you're a book festival organiser. You have the choice of coming up with a poster that is nice in a sort of generic way, or you can feature Chris Haughton's dog from Oh no, George! or Nadia Shireen's Billy and the Beast, or Melissa Castrillon's wonderful organic patterns. The first choice - the AI image - expresses disdain for illustrators (which they will note), and the second, besides tapping into a brand, lets you show off your connection with an amazing creator and highlight how your team support artists. If that's the mission of your festival, this absolutely makes sense: it makes you look good, draws extra attention to your festival, and taps into the artist's own fan base.
For writers, associating with real artists boosts your reputation. Illustrators scrutinise authors before they agree to take on their books. If you have credited artwork on your website (instead of uncredited, or AI), a top artist may be much more inclined to work with you, and to big up your project afterward.
But AI issue poses deeper questions about what it means to be a creative person. What can we make that a computer can't?
Over the past decade, I've seen social media turning so many of us into something that might as well be generated by AI. Illustrators copy online illustrators who have copied other online illustrators.
Our artwork isn't only about creating a certain style, it reflects ideas, and communicates how we see the world. And I'm not sure how much we really ARE seeing the world, and how much we're looking at it through the lens of other people, because that's what's acceptable in our bubble or peer group. It's easier to accept what people whom we admire say about an issue than to dig deep, seek uncomfortable truths and details, and draw something that freshly interprets what we've discovered. Social media works like ChatGPT; it aggregates us in a way that suppresses anomalies and outliers, and comes up with a convincing-sounding average blend. But that uniform result, just like ChatGPT, gets things wrong, and sounds boring. Behavioural science expert Rory Sutherland argued that if, using AI, 'you could produce someone who is the average of all your friends, you probably wouldn't like them very much and they'd be extremely dull'.
'But we just need to make a living', you could reasonably argue. 'If publishers are playing it safe, and I'm turning out the bland content they want, at least it pays the bills.' But this is unsustainable, what distinguishes us from AI? And why should anyone care if we disappear?
I don't think there's one solution to this problem, and perhaps therein lies its beauty. You need to find your own way, discover what makes you and your work unique.
Perhaps this involves stepping back from the computer for decent lengths of time, building off-line relationships with people, and learning how to make images that have nothing to do with anything going on in your online peer group. Try your hand at something that has nothing to do with a suggestion from an internet influencer, and don't worry about how it will look on social media. Explore it and revel in its unique quirks and eccentricities. For me, I've recently enjoyed drawing scenes of local community life; people I've been getting to know - they love seeing themselves in pictures, and I can draw them in very specific locations, with an awareness of their personalities and loveable quirks that a computer hasn't yet cracked.
Writer and illustrator Sarah McIntyre is currently creating the Adventuremice series with her co-author Philip Reeve and is trying to spend less time on the internet. And yet, there she is: jabberworks.co.uk
*Pictures Mean Business website generously created and funded by freelance artist Soni Speight. #PicturesMeanBusiness hashtag created during a discussion on Twitter with artist James Mayhew.