New tools like DALL-E, MidJourney and ChatGPT appear to be capable of rivalling human imagination. On the surface at least. What does this mean for our powers of creativity, a quality seen by many as one of the last domains where we humans have an advantage over artificial intelligence?
For the image accompanying this article, I asked DALL-E (a tool developed by OpenAI of ChatGPT fame) to generate a self-portrait of a woman working on a laptop, in the style of Van Gogh. The output is interesting I think, eye-catching even. But is it art?
Artists have responded scathingly to images like this. They point out that the outputs are derivative - generated by scraping the web and reproducing existing works - setting the scene for legal action preventing AI images to be commercialised. Their point? The algorithms that produce this art are undeniably clever, but the data used is largely generated by humans, and as humans we have rights when it comes to our creations. Fearing copyright action, many online image marketplace like Getty Images now ban the sale of AI generated images.
“The algorithms that produce this art are undeniably clever, but the data used is largely created by humans, and as humans we have rights when it comes to our creations.”
The prospect of legal strife aside, it’s hard to say what else the future holds for automated imagination. When do the benefits of innovation sufficiently counterbalance the harm they cause? As my friend Tom asked me over lunch yesterday - “were we better off before the internet?”
Will the robots put creatives out of business?
Historically humanity has found it hard to predict the actual long term impacts of new, disruptive technologies. Some fears are justified but the upsides can’t be ignored either. Robotic automation reduces the need for workers whilst making manufacturing both safer and more efficient for those that remain. The internet has made it easier for both knowledge and misinformation to spread.
What does this mean for automated creation of art, both written and visual? Intriguing as these tools are, I don’t think they’re as pivotal a piece of disruptive technology that many make them out to be, for now at least. For a start the murky ethics, copyright concerns and workforce exploitation that accompany their development can’t be ignored. Even if regulators find themselves, yet again, on the back foot they’ll eventually catch-up.
And we certainly shouldn’t call time on artistic endeavour.
As one art professor recently told Motherboard “Personally, I’ve been encouraging students to explore this. I think they should know what the tools are, what they’re capable of and maybe develop a personal vocabulary of how to use them”
That’s good advice for the rest of us too. Whatever your business or sector it’s important to understand how this new generation of data-powered AI tools might affect your area of work and see if there are ways you can use them, whilst also considering the harm they may cause.
Want to learn more?