Will your job soon be automated?
We are in the early tremors of a seismic shift in creative work, and in society as a whole. This month, in an interesting article about how AI could disrupt creative work, the Harvard Business Review claimed the AI revolution could herald the biggest social change since the creation of the printing press in 1439.
Anyone who has asked ChatGPT a question, or generated an image on DALL-E knows what it is already capable of. Even in these early days, it’s both terrifying and thrilling.
So how do we survive as creatives, in this brave new world? I’m no expert, but it’s important we start discussing this. So here are some initial thoughts. I’d love to know yours, too!
Understand AI’s limitations
The best article I’ve read so far about how the technology actually works is Cal Newport’s concise, clear explanation in the New Yorker, What kind of mind does ChatGPT have?
It’s worth reading, because it explains what a large language learning model actually is. But also what it is not. And as a Computer Science professor, Newport has more understanding of this than most. He explains why ChatGPT is not sentient. Nor will it grow to be.
It is brilliant at learning natural-sounding language, but It has no common sense. It often fails to understand the most ordinary and obvious things because it has no lived experience.
It can’t adapt to the situation at hand. Nor can it read the room, understand office politics or pick up the nuances of human communication that we tend to process so quickly that we’re often not even aware we’re doing it.
Being human is a superpower
“Much of what occurs in offices.. doesn’t involve the production of text,” explains Newport. “And even when knowledge workers do write, what they write often depends on industry expertise and an understanding of the personalities and processes that are specific to their workplace.
“Recently, I collaborated with some colleagues at my university on a carefully worded e-mail, clarifying a confusing point about our school’s faculty-hiring process, that had to be sent to exactly the right person in the dean’s office. There’s nothing in ChatGPT’s broad training that could have helped us accomplish this narrow task.”
Skills we’ll need survive the AI revolution
ChatGPT has no access to the internet, and its programmers stopped feeding in content in 2021 so there’s no point expecting it to know anything more recent than that. But even before that date, it has glaring gaps in its data.
I’ve asked it to tell me about several of my friends, both famous and not. Some it got with startling accuracy. But more far often, it delivered a cocktail of half-truths and misinformation with a confidence and plausibility that any politician would envy.
When I asked it about myself, it said I have written a book about club culture (true). But also a book about reality TV and The Ultimate Guide to Great Sex (not true). I’m happy to be credited with a 2005 Journalist of the Year award – but if I really did win this, no one ever bothered telling me.
AI can be a great research tool, a starting point. But don’t believe all it tells you. Be meticulous. Check your facts. Truth matters, and those who care about the details will stand out.
Protect your intellectual property
AI is trained by scraping content from the internet. And by ‘content’, I mean your work. And mine.
Getty Images are already suing the AI art tool Stable Diffusion for scraping watermarked images from its site, and it will be interesting to see how that plays out.
British organisations such as the Society of Authors and the Association of Photographers have useful advice on this, and are also campaigning for the law to catch up with the technology.
Wherever you live, there are probably local organisations doing the same. Support them. Be even more careful about the copyright agreements you sign. And do whatever you can to protect the work you make.
Prepare for disruption
The AI revolution is coming, ready or not. As the Harvard Business Review reminds us, the centuries that followed the spread of the printing press “featured rapid innovation, socio-political volatility, and economic disruption across a swathe of industries as the cost of acquiring knowledge and information fell precipitously.
“We are in the very early stages of the generative AI revolution. We expect the near future therefore to be more volatile than the recent past.”
This will be scary. But as always with disruption, there will also be new opportunities. Watch for them. And act quickly when you see them.
Get good at joining the dots
AI can search and collate information with a speed no human mind could hope to emulate. But we excel in deriving meaning from that information, curating what’s important, connecting seemingly unrelated ideas, making leaps in understanding.
Organising, structuring, innovating – handling the raw data and making new connections that only a human could make – will be key skills to nurture as AI develops.
Get good at making sideways leaps, new connections. At reading between the lines, lateral thinking and joining the dots. At editing, reworking, remixing, reimagining. These will be skills that machines can’t emulate.
Get comfortable talking to AI
The tech can only do what you ask it to do. So the better you frame your questions and prompts, the more useful its output will be.
Start now. Experiment, and play with the available tools. Prompt-engineering will soon be an in-demand skill. Those of us who are excellent at instructing AI will produce better quality work. And quality work will always be in demand.
Express yourself
A friend recently challenged me to read a blog post she’d written using ChatGPT and identify what the AI had created, and what was hers. She thought I’d struggle, but it wasn’t hard.
I was a magazine editor for many years. I’m used to looking at words in forensic detail. Sentences created by AI that at first seem passable tend to fall apart on closer examination. You soon realise they’re not actually saying much of substance. Let alone saying it with any style.
My friend, however, is a fine writer who rarely waffles or repeats herself. She has a strong, recognisable voice and a point of view that is always interesting. It wasn’t hard to tell the difference between her words, and the ones culled from ChatGPT. Most of the bits I got wrong were where she’d adjusted her own style to fit more seamlessly with the AI content she was weaving in.
AI can emulate jokes, but it isn’t witty. It can’t come up with startling new metaphors. Or new ideas.
Most of all, it has no strong opinions, no personality, no unique point of view. The same is true of AI generating visual content, or music. There’s often something off in the images. Faces are wrong, a hand is positioned oddly. AI music creates atmosphere, but it’s never going to capture a mood and express exactly how we’re all feeling, in the way a great song does.
So let you be uniquely, gloriously you. With all your quirks and flaws. No machine can do that.
“Algorithms don’t feel. Data doesn’t suffer.”
When asked about some lyrics ChatGPT produced in his style, Nick Cave put it beautifully in The Red Hand Files. “It could perhaps in time create a song that is, on the surface, indistinguishable from an original, but it will always be a replication, a kind of burlesque.
“Songs arise out of suffering, by which I mean they are predicated upon the complex, internal human struggle of creation and, well, as far as I know, algorithms don’t feel. Data doesn’t suffer.
“ChatGPT has no inner being, it has been nowhere, it has endured nothing, it has not had the audacity to reach beyond its limitations, and hence it doesn’t have the capacity for a shared transcendent experience, as it has no limitations from which to transcend.”
AI is can imitate a style, but it can’t create one. Nor can it layer in the meanings, associations, symbolism, cultural references that elevate any art from good to great.
It’s a tool, and I’m sure there will be artists who use it to create interesting work. But to innovate, you need a human directing and shaping its output.
We create meaning – and value
In April 2023, someone paid £209,000 at auction for a three-piece polyester suit with flared trousers, bought in a Brooklyn menswear store in the mid-70s. This seems illogical, unless I also tell you the suit is white and was worn by John Travolta in Saturday Night Fever.
We are willing to pay more for objects with perceived value. But that value can only be added by a human touch.
The mass production of the 20th century made us treasure the bespoke, the unique, the hand-made and imperfect even more. I think AI will perform a similar function in the 21st century, weeding out bland, vanilla creators but elevating those with talent and the patience to work at their art or craft.
I can see how AI will easily replace content farms, click-bait producers, bad writing, derivative art, off-the-peg design, bland opinions. lift muzak. What will become even more valuable now is work crafted with love and care; services delivered in a personal way that make us feel seen and understood; creators sharing their flaws and emotions, their suffering and their triumph, their weirdness and their wonder.
Without human creativity, our culture will ossify and die. And AI will have no fresh inputs to train on, no new work to imitate, no new ideas to regurgitate.
AI or human: which do you prefer?
I started work on this article with a question: how can creative professionals thrive, as generative AI become tools more available?
You’ve just read my thoughts on this. If you’re interested, you can also read how ChatGPT answered the same question.
Here’s the article it generated: Navigating the AI era. It’s illustrated with another image created by Dall-E. I’d love to know how you think it compares with this one.
What do you think?