AI-generated art illustrates another problem with technology | John Naughton


IIt all started with the title of an entry in Charlie Warzel’s book Galaxy Brain Newsletter in the Atlantic: “Where does Alex Jones go from here?” It’s an interesting question because Jones is such an extreme internet troll that he makes Donald Trump look like Spinoza. For many years, he turned a radio talk show and website into a comfortable multi-million dollar business peddling nonsense, conspiracy theories, lies and weird wares to a huge tribe of adherents. And until August 4, he got away with it. That day, however, he lost an epic libel case brought against him by parents of children who died in the 2012 Sandy Hook massacre – a tragedy he had constantly ridiculed as a staged hoax; a Texas jury decided he should pay nearly $50 million in damages for publishing this sadistic nonsense.

Warzel’s newsletter consisted of an interview with someone who had worked for the Jones media empire in its heyday, and as such was interesting. But what really caught my attention was the striking illustration that ran the room. It showed a cartoonish image of a disheveled Jones in a sort of cave surrounded by papers, banknotes, prescriptions and other types of documents. Pretty good, I thought, then inspected the caption to see who the artist was. The answer: “Art of AI by Midjourney”.

Ah! Midjourney is a research lab and also the name of its program that creates images from textual descriptions using a machine learning system similar to that of OpenAI. Dall-E system. So someone on the Atlantic simply typed “Alex Jones in an American office under fluorescent lights” into a text box and – bingo! – the illustration that caught my attention was one of the images it generated.

It turns out that the Atlantic is not the only established publication in which the work of the Midjourney tool has appeared. The normally seated Economistfor example, deployed it recently to produce its June 11 cover. This is important because it illustrates how quickly digital technologies can move from cutting-edge to commoditization. And in doing so, new fears and new hopes are rapidly emerging.

Dall-E (the name is a geeky combination of the Pixar character Wall-E and Salvador Dalí) is derived from OpenAI pioneer GPT language models, which may generate vaguely plausible English text. Dall-E essentially swaps pixels for text and was trained on 400 million pairs of images with text captions that were “scraped” from the internet. (The carbon footprint of the calculation involved in this process is inadmissible, but that’s for another day.)

When GPT-3 emerged, it sparked a new episode of the “increase v replacement” debate. Was technology just the thin edge of a grim corner? GPT-3 could be used to “write” boring but useful texts – stock market reports, for example – but it could also generate harmful and seemingly credible misinformation that would pass through the moderation systems of social media platforms. It could be used for increase capacities of busy and overworked journalists or for do without it entirely. Etc.

In this case, however, some of the steam has come out of the GPT-3 controversy (though not out of the question of the environmental costs of such extravagant computing). No matter how many skeptics and critics might ridicule human hacks, the crooked wood of humanity will continue to outsmart simple machines for the foreseeable future. Journalism schools can relax.

Dall-E might turn out to be a less straightforward case, however. As with GPT-3, its appearance has sparked a lot of interest, perhaps because while most people can write text, many of us can’t draw to save our lives. So having a tool that could allow us to overcome this handicap would be a godsend. You could, for example, ask for a portrait of Shrek in the style of the Mona Lisa or Jane Austen as an astronaut and again he would do his best. This can therefore be seen as a welcome increase in human capacities.

But there is also the question of “replacement”. Turns out it was Warzel himself who used Midjourney’s bot to create an illustration rather than getting one from a copyrighted image bank or having an artist create an image. Big mistake: An artist spotted the caption and tweeted that he was shocked that a national magazine like the Atlantic used a computer program to illustrate stories instead of paying an artist to do this work, thus giving other publications the idea to do the same. Before you can say “IA”, Warzel He found himself playing the villain in a viral tweetstorm. Which was painful for him, but perhaps also a salutary warning that publishers who give work to machines rather than creators deserve all they get.

what i read

Good functioning
Electric vehicles are much more energy efficient than internal combustion vehicles a sobering summary of the Yale Climate Connections project.

Get better
The efficiency movement is a wonderful essay by Rob Miller how all modern societies have been shaped by their cult of efficiency.

Biological clock
The Nautilus site offers an exciting article on the evolving mysteries of menopause.

Previous The Tolino Epos 3 eReader is now available
Next K-boom! South Korean art and high culture come to Britain | South Korea