A Report from the Frontline of AI in Production

With the emergence of generative AI now dominating conversations across the industry, we can start to ask; How is AI being used in production, not how will it be?

And if you ask global production house The Sweetshop, they’ll tell you the answer: “for almost everything”.

Increasingly, it feels as though the industry is approaching some kind of precipice; or a point of no-return. As we start to think in terms of a pre-AI and post-AI timeline, LBB set out to discover how companies were building the technology into their workflows today.

To do so, Tom Loudon sat down with some of the key creators and directors at The Sweetshop, including digital artist and futurist Tomas Roope, film and game director Joshua (JH) Walker, digital artist Jacqui Kenny, and Sweetshop co-CEO, founding partner, and director Melanie Bridge…

LBB> What are you currently working on using Midjourney, or any other AI?

Tom> We currently work across a number of platforms including DALL.E 2, Stable Diffusion, Midjourney, and Runway.

Jacqui> Working on these AIs amplified my feeling when I worked on an art project using Google Street View, that the worlds I was escaping into were not representations of our world, but distinct other worlds that technology was revealing.

From this we started working on an art and film concept called Argleton Lane. Argleton was a town that appeared on Google Maps in the 2000’s and is believed to be in the tradition of trap streets, where map makers added phantom data to maps to protect their copyright. The idea of Argleton seemed symbolic of the myriad of digital simulations of the world created by big companies and agencies, and the representations of “us” that inhabit them. To date we have launched two NFT collections around the concept of Argleton Lane. The first using GAN technologies and the second using the first generation of the generative models.

Melanie> We look at AI as simply another tool in the toolbox. Every day at The Sweetshop, most of us are doing some sort of work in Midjourney, chat GPT, or some other AI programme to help visualise projects or expedite the writing process. In particular it’s an incredible resource for directors’ treatments. AI has got to the stage now where with the right prompts the result is virtually photographic, which makes it quick and easy to express an often hard-to-explain idea or concept, literally getting it out of your head and onto the screen within a couple of minutes.

Joshua> I’m particularly drawn to crafting human images and narratives in Midjourney. I love pushing myself to see how closely I can align my creations with the emotions and ideas I envision. I now incorporate AI into my daily creative routine.

LBB> How do you believe the emerging AI space will impact your work?

Tom> My work has always focused on emerging tech and how it intersects with culture at scale. Generative AI is just another - though incredibly significant - technological evolution. I have been working on AI projects since 2017 and after leaving Google in 2019, it has been my sole creative focus.

Jacqui> Generative AI has been a natural evolution in my post-photography practice. It has meant that I have had to learn how to effectively translate my imagination into English words, something that has become more powerful than ever. Generative AI is the latest phase of a revolution in creativity, summed up in Brian Eno’s statement, “composers are becoming more like gardeners than architects”.

I’m still using my original street view images as seeds, combined with text prompts in DALL.E and Midjourney to create my current collection.

Joshua> I have discovered a creative freedom of expression I’ve never had, with tools that allow me to bring ideas to life without the logistics and costs of a traditional production workflow.

But with Midjourney, I can be my own concept artist and create designs that show precisely what I want, including all cinematic elements- tone, angle, lighting, lensing, and characters—all in a single contextual frame.

LBB> Could AI programs like Midjourney impact gaming conception, advertising, and film in a negative way?

Tom> A few years ago I read a research report titled ‘Misery is not Miserly’ that showed that "sad and self-focused individuals" spend more, and have since wondered whether the negative impact of social media is down to unintended effects of AIs relentlessly pursuing its sales goals. Adding generative AI to the mix will only worsen things, unless we do something about it.

Joshua> There's something special about human creativity, like when you're just lying under a tree, watching the clouds drift by. AI has no soul, so when we allow it to control our imagination, I fear that we’ll lose ours. If we lean too heavily on AI-created content, we might end up with a bland sameness in our ideas.

LBB> How can visual artists use Midjourney in advertising?

Tom> Directed by prompts that dynamically combine brand messaging with individual user insights will create personalised marketing messaging in real-time, at scale. The creative challenge, as ever, is how to transcend the functional to achieve the exceptional.

Jacqui> You can now train and prompt characters and products to generate hyper-personalised imagery that can be tailored at a one-to-one level.

We’re only just beginning to see the potential, and brands are very excited but also concerned about the legal implications of using AI, especially AI art. Once there is a clear legal position on creating work with these technologies, there is going to be a proliferation of generative AI advertising.

AI is an incredible creative collaborator, so the possibilities are endless.

Joshua> As a filmmaker who's all about emotions and visuals, I'm always looking for ways to grab my audience's attention. Generative AI has been a game changer for me; it just takes things to a new level. It's so much faster and more accurate in bringing my ideas to life and lets me dive deep into my creative process.

LBB> What aspects of AI impact your work but are overlooked in the media?

Tom> Generative AI has moved us into an era of infinite interns, all of whom have decent illustration, writing, photographic and photoshop skills who can travel to any time and place. This has meant that human curation has become critical, as the direction and assessment of this team of infinite interns are far more manageable than the output from the monkeys.

Melanie> I’ve been listening to many AI experts, and all seem to agree that the systems must be made more transparent to protect people from disingenuous AI trickery (whether video, art, or communications), and it should be a legal obligation to label them as such by the programs (or people) that create them.

My greatest bugbear with the current state of AI is the ability to use artists' names in prompts to create artwork in their style. Why is it ethical to provide a mechanism for any person to effortlessly and openly plagiarise an artist’s hard-earned style? This would be easy to control by banning people's names in a similar way that ‘rude’ words are banned from prompts. This would also stop people from using the likeness of any person, which benefits no one.

Joshua> Breaking free from geographical limits and teaming up with people worldwide is a game-changer when you use AI-driven tools like Midjourney, making brainstorming and pre-production bits way faster. It's incredible how we can create entire worlds together through the collective imagination of artists worldwide.

LBB> You’ve been a part of some big projects - how do you maintain the drive and inspiration needed for your work, knowing that Midjourney could potentially revolutionise your processes?

Jacqui> Midjourney and other generative AI platforms have become a natural progression of my work. Most of the projects I have worked on over the last six years are rooted in the idea of travelling without travelling.

Melanie> Currently, we feel more excited than worried about what is to come. At the highest end of the market, where craft and authenticity are key, we believe AI will affect us less negatively. At the two most recent award shows I’ve been to, I’ve been heartened to analyse the winning entries and realise that there were not many (if any) that could have been created using AI. Authentic human stories with real people and great ideas are still winning nearly every time and everywhere, thank goodness.

Joshua> I've always believed in staying true to my passion for storytelling and valuing my unique contribution as a human artist. Midjourney is a powerful tool, but its impact depends on the person using it. I'm all about focusing on personal growth and honing my skills to keep pushing the boundaries of my work and stay ahead of the curve.

So, although AI won't ever replace the passion and human connection that drives my work, I'm hopeful it can enhance my existing abilities and lead to even more amazing opportunities.

LBB> What is something you always aim to achieve within your work, and how does the introduction of tools like Midjourney affect this?

Jacqui> Generative AI platforms are freeing me to have a far more direct relationship with my work. I have always loved exploring ideas and imagery but have never felt comfortable with real-world shoots. I now have an unmediated relationship with bringing my ideas to life, which is incredible. Curation is still critical as though the output is incredibly real. It is only exceptional by accident, in exactly the same way as my street view project where I had to sift through tens of thousands of images to discover the magic.

Joshua> The most important thing in my work is to evoke genuine emotions and connect deeply with the audience. I believe that cinematic filmmaking combined with authentic performances is the best way to achieve that.

AI has forever changed the way we present our work. I remember when we used to show our treatments to agencies using these huge boards that we carried in special folders. Half the time, the pictures would fall off while we were on our way to the presentation!

Article Images from Artist Jacqui Kenny

The NFT Images featured in this article were part of the group photography show, Post Photographic Perspectives II. This show featured 10 artists, and 1200 images in total—everything sold out in about 24 hours.

This article was first published by LBB Online on 24 July 2023.