Rethinking AI images on The Cybershow


Figure 1: "Welcome to the machine"

We've been using generative AI since the launch of The Cybershow but have recently changed our minds about it. It's a sad decision, because we love the technology and the aesthetic it gives the site. But something feels wrong.

Our brave new world

It seemed natural to just dive-in with an emerging technology and explore it. The site gained a lovely modern and fresh feel from the get-go.

It also served to signify that we are very pro-technology and generally optimistic - but in a world where criticism of monopoly, reckless engineering and technological abuse are mischievously characterised as "Luddism". It's hard to accuse a group of being anti-progressive on AI when the front page is literally a gallery from the latest algorithms.

From the start we eschewed "photo-realism", electing instead for stylised output in various genres from pop-art to dank meme. We were also one of the first regular users to voluntarily label AI artwork as such. It always seemed problematic to me as a designer that there was nobody to credit, so at our launch party we tagged the videos for YouTube with a notice and details of the tech used. That is now a standard practice Google are advising. We also noticed that within a few months the NCSC's site sprouted strangely similar aesthetics to our own. I'd say we've enjoyed being trend-setters and it's been a fun ride.

Second thoughts

On a practical level, it's a very useful human search aid to associate each post with a recognisable image. We wanted to avoid time consuming visual work in order to concentrate on writing and audio production.

However, there is a certain pleasure in thinking about visual signals and what we want to communicate as a manual process. It seems faster to experiment with pen and paper. Besides, my 'shopping skills (actually Gimp and Inkscape) are not too shabby having been a lurking B3tard back in the millennial days, and I still love that self-mockingly rough-cut meme style of the 2000 - 2010 decade.

Doubts emerged late last year after Helen battled with many of the generative platforms to get less racist and gendered cultural assumptions. We even had some ideas for an episode about baked bias, but other podcasters picked up on that and did a fine job of investigating and explicating.

Though, maybe more is still to be said. With time I've noticed the "guardrails" are staring to close in like a pack of dogs. The tools seem ever less willing to output edgy ideas critical of corporate gangsters. That feels like a direct impingement on visual art culture. Much like most of the now enshitified internet there seems to be an built-in aversion to humour, and for that matter to hope, love or faith in the future of humaity. The "five giant websites filled with screenshots of text from the other four" are devoid of anything human.

Like the companies that make them, commercial AI tools seem to have blind-spots around irony, juxtaposition and irreverence. They have no chutzpah. Perhaps we are just bumping into the limits of machine creativity in its current iteration. Or maybe there's a "directing mind", biasing output toward tepid, mediocre "acceptability". That's not us!

As Schneier writes;

"The increasingly centralized control of AI is an ominous sign. When tech billionaires and corporations steer AI, we get AI that tends to reflect the interests of tech billionaires and corporations, instead of the public."

Of course we have the technical chops to put a few high end graphics cards in a rack and run our own uncensored models. But is that a road we want to go down? Do we want to adopt the technology of the enemy when it might turn out to be their greatest weakness, and our humanity our greatest strength?

The Cybershow is growing to include diverse contributors, and is as much driven by artists as hackers and engineers. So we need to ask;

  • What signal does using AI imagery send about the show?
  • Are we being hypocritical to our own ethics?

If you lie down with dogs, you get up with fleas

Perhaps the real problem is with association. Much as we love generative AI as a visual technology for one specific task we categorically do not use it in writing, or for video and sound. All of us are involved in original research, creating prose, poetry, writing and recording the music. We manually produce each episode using open source tools like Ardour and KDEnLive. That's the pleasure of the art and craft of our show.

But what does it say to followers when the visual site aesthetic trumpets "AI is cool!"? Does it create a suspicion that more, maybe all of the other content is machine made?

Another fact is that the cultural tide is turning around AI. People generally dislike and fear it. Like social media, AI is concentrating still more abusive power into the hands of the incumbent tech oligopoly. Its potential as an equalising force seems already dead, and it seems ever more like a technology aligned with all the things we are against.

A spring clean

So starting with this Spring Equinox we'll not be creating any more pure AI images for the site. It's been a difficult decision. We'll also slowly be removing and replacing some of the older images with hand illustrated or hand-crafted digital montage type works.

For now, we will retain AI visuals for two uses;

  • To quickly brainstorm some ideas and see how they might work
  • for one-shot memes for social-media

Then we can say with confidence "this site is 100% human created".

Date: 22 March 2024

Author: Dr. Andy Farnell

Created: 2024-03-22 Fri 19:33