Sci-fi nightmares of robot take-over often start with the working-class jobs.
It’s almost always assumed that the farm hands, assembly line workers and shop assistants will become robots, or apps. Avowed techno-optimists might even push for this to happen faster: it’s for the best that these ‘low skilled’ and often dangerous jobs get automated, right?
Then maybe automation will come for the more wearisome back office work of the professions - the paralegals, bank clerks… Then who knows? But maybe everyone will be free to pursue the ‘softer’ world of human creative expression. We might all be like Joaquin Phoenix's character in ‘Her’ - writing perfect love letters to order, while all the hard, boring stuff is done by robots.
That story has been upended by the arrival of the likes of Dall E, and Chat GPT. The advances in generative adversarial networks have hit an inflection point in the last 12 months that’s led to many speculating about the automation wave hitting ‘the creatives, not the ‘low skilled’. Could the need for writers writing or illustrators illustrating or photographers photographing be going up in smoke at the hands of the very expensive data engineers at Open AI?
Even more dramatically, does this mean Art - or at least humans making it - is over?
Art has changed a lot in the last few thousand years. But it’s still here
It’s almost obligatory to show a cave painting at this point, so indulge me. Creative expression has been with us since we were us - perhaps even before various other key technologies like fire and cutting tools emerged.
And the human brain, and our cultures have continuously found new forms, media and ways to do it, even when times are tough. The arrival of new means - technologies, ideas and techniques - has if anything led to a proliferation of more creativity. Not less.
Painting for many years was the core ‘technology’ for visual expression in the western world, but this was shaken up in the 19th Century with the arrival of photography and film. Without the need to sit with a portrait artist for 10 hours to capture your likeness, surely painting was dead? Photography could just neatly supplant all the uses of painting leaving the brush wielders to look for the next thing. Well, of course it didn’t.
Firstly, it turned out that just having a camera wasn’t enough to make you a great photographer. The skill, understanding and manipulation of that technology continues to underpin the difference between amateur shots and the work of a professional. Even more still the distinction between a professional and a world renowned artist using it as a medium.
Secondly, there wasn’t a neat displacement where photographs entirely swapped in where paintings made sense before. Yes, getting a portrait painted is now largely a niche for the royals, pompous tryhards and our fur babies to inhabit. But painting is still here. Artists spent the last 100+ years expressing novel work and broader movements within it. If anything, its true strengths and breadth have been surfaced by the removal of the ‘need’ to do what photographs do well.
Thirdly, there was a cascade of things you couldn’t have predicted that photography and film opened up well beyond the bounds of painting. Magazines shifted dramatically to more visual technicolor medium in their own right, with cascading effects across consumer products, advertising, human self-perception… and yes Art. Cinema, Hollywood and multiplexes emerged, alongside film as an artform too. And 150 years later, innovation in computer vision, diagnostics and automation is all underpinned by the ability to capture light through a lens.
Generative AI as part of the toolkit, not the only show in town
So what might Generative AI do in the world of creative expression, given all the above.
Well - there certainly is likely to be some disruption. If you were making money from churning out fairly identikit web copy to boost SEO, you’re probably already thinking about your next move. But if photography is anything to go by, this isn’t going to be a simple transition. We might all be able to have some fun with Chat GPT and DallE. but the emergence of ‘prompt engineering’ as a skill set and an idea suggests we are still working out how to master these complex algorithms and direct them effectively.
Secondly, this isn’t going to be a neat switcheroo. You won’t swap in all writers for Chat GPT, but Chat GPT might take some of the tasks off of what we call writers plates. Or it might be more successful in making us write those meeting notes like Heamingway but stay out of the novellas. Similarly, we should expect a counter reaction in the actual medium of writing; I wouldn’t be surprised if ‘writers’ of all hues moved actively away from the Google enforced bland and reading age minimising clarity.
Thirdly, we’re already seeing the proliferation of artists tinkering with these algorithms and tools to generate new forms of self-expression. I’ve particularly enjoyed the work of Holly Herndon lately - and she’s got a great podcast that often explores the emergent possible with other creators. It’s early days, but it’s often in the hands of these more radical and explorative minds that the frontiers of a new technology emerge.
What does this all mean if you, your team, or your business rely on making content or creativity more generally?
There is a lot happening all at once, at different layers of the product stack around generative AI. Some of it feels immediately magical. But closer looks tend to reveal true limitations and strengths. What’s the best way to keep up?
First, keep track of the product frontier - and potential productivity gains and creative outlets as they emerge. There’s practically a new product concept each hour shipping on the back of these models. Check them out, try them out. In some cases, it’s clear we’re looking at exciting features of the future, in others, the limitations of the technology are exposed.
Second, run a lean futures workshop with your team. What are the axes of uncertainty across the technology landscape today (AI and beyond)? What shifts in the lives and expectations of your customers are possible? And which parts of your business are likely to need to move to meet them?
Finally, get making at the fringes. The open nature of these tools, and the commercial opportunity that the likes of Google and Microsoft sense around enabling their application means you don’t need a fully functioning data science lab to make use of them. Even low-code automation tools like Zapier are starting to enable insertion of OpenAI into workflows - a perfect way to start making and putting these capabilities into people’s hands rapidly.
We’re just getting started with exploring what this new toolkit means. Get in touch if you’d like to talk, or start tinkering with the future of it inside your org