P.png

Hello, we’re Pellicle

We’re your favourite drinks magazine and podcast, all about beer, cider, wine, pubs and more. Reader supported, proudly free to read.

Do Androids Dream of Electric Pints? — On AI Art in the Beer Industry

Do Androids Dream of Electric Pints? — On AI Art in the Beer Industry

There isn’t a commercial brewery out there that doesn’t, on some level, have a passion for the beer they produce, the ingredients they use, or the techniques they employ. Why, then, when it comes to packaging, are an increasing number of businesses adopting the often soulless output of modern generative AI software—and what impact will this have?

While the emergence of adaptable, user-friendly AI technology brings immense potential, it also raises serious ethical questions. A dystopian Willy Wonka experience that broke the internet and the recent, morbidly fascinating Royal Family discourse has drawn the use of AI into the mainstream media. It's no surprise, then, that it’s already crept into the beer industry. AI-assisted recipes debuted in the form of Neural Network from Australia's Modus Brewing in May 2023, and AI-PA from Suffolk’s Nethergate Brewery in January 2024. But perhaps the biggest impact of AI in beer and brewing will be how it’s utilised within the design process.

It's easy to understand the allure of “generative AI”; bespoke images, created seemingly from scratch by prompting software, requiring no more human effort than it takes to use a search engine. Proponents of the technology claim that it democratises creativity—that those unable to turn ideas into finished works now have that ability, bypassing years of training and experience accrued by professional writers, illustrators, designers, photographers, and videographers. These systems are “trained’” on vast swathes of data; they have been fed billions of images (or lines of text), giving them the ability to emulate existing visual or literary styles. They can generate in minutes something that may have taken hours, days or weeks for a human to produce.

“Broadly, I’m optimistic about the opportunities that [it] could unlock for us as people trying to be creative in small businesses and challenger brands,” says Andy Nowlan, marketing manager at Finchampstead-based brewery Siren Craft Brew. “I'm trying to stay open-minded.”

With scores of ambitious, mission-driven “challenger” brands (essentially, those between the niche and the mainstream) competing against established household-name brewers for space on bar taps and supermarket shelves, a strong visual presence is more crucial than ever. Over 1000 new beers are released every year in the UK alone. That requires a lot of design, illustration, and copywriting, and somehow it all needs to be distinctive, engaging, and preferably not blow the marketing budget. Considering how hostile the trading conditions within the beer industry are at present it's unsurprising that some businesses have turned to AI, either leveraging the technology for PR, or utilising AI on label art.


“AI gives us a platform to still be creative, and express our desired look in a cost effective way.”
— Cameron Brown, Turning Point Brew Co.

“We’ve collaborated with lots of different artists over the five years we’ve been a brewery, but the majority of our recent labels were created by a freelance designer who we’ve worked with since 2021,” Richard Wardrop, marketing manager at Edinburgh’s Vault City brewery tells me. They recently used prompt-generated AI images (albeit edited by a real person before their eventual use) for the release of their trio of Iron Brew sours.

“They created some initial concepts for our Iron Brew launch and although they looked great, we wanted to differentiate more from our standard rotational releases, Richard says. “[They] shared some more ideas and among them was an option created using AI.”

Knaresborough's Turning Point Brew Co are among the most prolific users of AI imagery in the UK beer industry. After their long-term collaborator stopped producing illustrations for the brewery, they took the work in-house. Owner Cameron Brown began designing the labels himself, using an AI engine called Midjourney to generate images.

“The problem we have is making thirty to forty new beers a year, and wanting each one to look and feel a certain way; the cost of this on our [production] scale is unrealistic,” Cameron tells me. “AI gives us a platform to still be creative, and express our desired look in a cost effective way. We've always had a sci-fi/futuristic look and feel to our branding, and I'm able to get great results using Midjourney in a wide range of styles.”

Despite being a relatively young industry, AI is growing at a pace that could be described as frightening. Valuations vary wildly, with several suggesting the industry could currently be worth over £400 billion globally, and expected to quadruple within the next decade, with the technology providing what’s considered by some to be a cheaper alternative to hiring staff or commissioning agencies.

Illustrations by Adam Menzies

“There are major UK breweries using AI designs, and plenty more businesses who may use it for ideation or inspiration,” Richard says. “The brewing industry is under extreme pressure with rising material and production costs, so there could be more businesses experimenting with AI designs and beyond to save money, or increase productivity.”

***

There was once a time when many at the cutting edge of the beer industry considered “productivity” to be a dirty word. In recent years, increased demand coupled with rocketing costs have forced even the most artisanal of brewers to reckon with their efficiencies, whether they discuss this publicly or not. Few of the options by which brewers might attempt to balance the books, such as using cheaper ingredients or making personnel cuts, are viewed favourably—so the appeal of affordable, rapidly-generated images is understandable. But despite outward appearances, AI does come at a high price—one being paid by the creative industry.

Generative AI software is “trained” almost exclusively using intellectual property, without the copyright holders’ permission. Although difficult to pursue, this still constitutes copyright theft—a civil wrong, and under certain circumstances, a criminal offence. Already, numerous notable legal cases are underway; in 2023 The New York Times and Getty Images filed lawsuits against OpenAI and Stability AI respectively. It has also been confirmed via leaked internal comms that Midjourney is designed to emulate selected “styles”—meaning the work of specific artists, who have not been remunerated for the exploitation of their work, and may not ever be paid back. Researchers have demonstrated that several generative apps consistently regurgitate protected images as well as written copy.

There's a reason the technology’s detractors prefer terminology like “scrapeware” over “artificial intelligence,” given the software is not actually sentient, and in the case of image generation, can’t perceive its own output. It merely pieces together elements of billions of other images, like a waiter stealing individual morsels of food from competing restaurants and presenting them all on a plate as an original dish. It’s a technology that is becoming increasingly divisive—particularly with truly chilling advances coming thick and fast, such as those enabling the convincing imitation of a person’s speech patterns and facial movements.

Many hope that the law will soon catch up with AI proponents’ ethically cavalier advocates. Cameron, however, shrugs off concerns; he doesn't plan to go back to human-made illustrations unless Turning Point reduces its annual release count significantly—and it appears most of his customers share his nonchalance.


“There’s something about the packaging on a physical product like a beer bottle or can that holds meaning.”
— Andy Nowlan, Siren Craft Brew

“I think AI is still in its infancy, and there's a lot of varying opinions being formed in real time,” he says. “We share some of those concerns too, in the wider context. We might get the odd ‘not another AI’ on Untappd [...] but [it] rarely comes up with our trade customers, if ever.”

While retailers may not show concern (understandably, given the more pressing concerns facing retailers) many consumers consciously support small businesses as an ethical choice, and can be more vocal with their discontent. Upon the release of their Iron Brew sours, Vault City faced some backlash.

“Overall [the response] was more positive than negative, but we listened carefully to the feedback, especially from artists, designers and illustrators,” Richard says. “A lot of people liked the characters and the departure from our standard label designs. However, we also had comments from people saying they were disappointed we’d used AI, and even some saying they wanted to boycott the brewery.”

“We’ve worked with countless designers and illustrators over the years, all paid fairly, but even using AI once […] had people questioning whether we valued art. It’s a much more contentious issue than we anticipated, and with the feedback we aren’t likely to go down this route again for new designs,” he adds.

***

In a January 2024 survey by the Society of Authors, a quarter of the illustrators who responded claimed to have lost work to generative AI, while over a third have allegedly seen a decrease in the value of their income. 95% of respondents called for more safeguards and regulation of AI software.

Drew Millward, an illustrator based in Yorkshire’s Aire Valley, has worked with many beloved UK bands, festivals, restaurants, bars, and breweries. His psychedelic, highly-detailed style is engaging, distinctive, and to the untrained artist, difficult to replicate—exactly the kind of work generative AI is primed to exploit.

“It’s incredibly difficult to predict such a fast moving area, but, at present there is absolutely no ethical way of using the generative tools that are out there,” Drew tells me. “It’s built on the work of artists, pulled from every corner of the internet.

“I have real concerns about the ethical use of the tools, which are used to essentially create work of specific artists. That seems to me to be in the realms of plagiarism, which is a notoriously tricky legal area anyway, but made even more so when you factor in a solely technological generative approach.”

As a designer myself, I know designing beer labels is by no means hard graft, but it’s certainly not easy—it comes with its own set of cognitive challenges: Will these colours affect the consumers’ perception of flavour? Is it worth stripping out some detail for a marginal increase in legibility? Does this illustration carry any cultural cues that might make it a poor choice for the intended market? To the designer, every beer label embodies a set of decisions  as distinctive as those the recipe inherits from the brewer.

"For me there’s something about the packaging on a physical product like a beer bottle or can that holds meaning,” Siren’s Andy Nowlan tells me. “Somehow, when you know first-hand the amount of people, processes, hard work and time involved to get that beer onto the shelf it’s sat on, it’s a motivation to put the same time, effort and consideration into the branding.”


“A present there is absolutely no ethical way of using the generative tools that are out there.”
— Drew Millward, Artist

“Trying to find shortcuts just doesn’t sit right, however silly that might seem and whether or not the attempt at authenticity really makes it through to the customer,” he adds.

***

The future of the creative industry will undoubtedly have been impacted by the rise of generative AI—but there are still reasons to be positive. The first wave of copyright theft required for the software to function in its current form has already happened, and while tech companies may claim the data was ethically sourced, former employees are speaking out and confirming they knew it wasn’t, and used it anyway. It may be too late to reclaim what has been stolen—but artists are fighting back with digital tools like Nightshade, which acts like a poison, crippling AI’s ability to interpret images, thus protecting the work from being exploited for data sets in future.

Beyond copyright theft, artists’ other major concern—loss of work—remains. Again though, there is still hope. AI proponents claim artists need to evolve—but not every evolutionary step is a positive one. While designers can edit their files, AI has to start over every time. Illustrators can remove specific details from their work; AI can’t perceive what the details are. Writers develop an ability to deal with criticism and feedback; something the new wave of “prompt engineers” seems woefully incapable of handling. It is often said of commercial design that you can pick two attributes—fast, cheap or good— but never all three. AI appears to fall into the same trap.

Will we see AI replace the work of designers, illustrators and writers in the beer industry? Hopefully, those who hold care, passion and quality in high regard, will ensure not.

“I think it will take some hindsight to really appreciate the impact [of AI] but, if I’m honest, I’m fairly optimistic,” Drew says. “[AI is] a cheap, quick and unsatisfactory solution to a complex problem. The ‘problem’ being; how to make things look nice. Ultimately, that’s what art and design boils down to and, certainly from the examples I’ve seen, AI cannot generate that.”

“Aside from the ethics, the work just looks fucking awful.”

Power Fraus and Purple Fizz — In Conversation with Alanna LaGamba of Vin de LaGamba

Power Fraus and Purple Fizz — In Conversation with Alanna LaGamba of Vin de LaGamba

David Bailey's Just Beer Things #9 — Head!

David Bailey's Just Beer Things #9 — Head!

0