2022 was definitely a year for AI. Image generation in particular has taken off somewhat spectacularly this year, although we’ve also seen a lot of progress in the broader scope of machine learning. In fact, it has given us a lot to think about in the games industry. And while many will herald it as something that could solve all our problems, others have taken a slightly less welcoming approach.
As with any advancement, there is a fair amount of controversy involved. Beyond the iRobot-fuelled fear that AI will one day rise and take us hostage, there’s the (only slightly) less-worrying potential that it could displace countless artists, and it seems the process has already begun.
While rules have been put in place to prevent things from getting too out of hand, we’ve seen a fervent discourse erupt on the internet about the myriad of AI phenomena that are trending this year, even protests. But is it all bad?
With so many potential applications, it’s hard to ignore in any industry, and since it can learn much faster than us humans, I’m confident it will become even more prevalent in the years to come. Even as I write this, our robot overlords are no doubt plotting our demise, but AI has given us so much to be thankful for.
Like the relentless yet transformative march of the Roman legions, AI has permeated our culture for better or for worse. Now is the time to look back at how far AI has come this year and see how screwed up we are.
Robot overlords
A year of AI coverage started modestly for us. When Daniel West gave us the AI controlled Lego sorting machine (opens in new tab), he couldn’t have better prepared us for an explosion of artificially intelligent robots. And while we were all dreaming of the many organizational improvements AI could bring to our daily lives (just me?), MIT was accelerating learning by robots (opens in new tab) in incredible ways.
We even saw them train pizza making robots (opens in new tab) to pave the way for highly complex artificial intelligence – learning to manipulate deformable objects that change as they form has not been an easy task for AI robots to learn, but it all adds up to potential applications in for example healthcare. And it’s not just pizza that’s getting the AI treatment this year.
Boston Dynamics recently upgraded its artificially intelligent robot dog Spot (which semiconductor manufacturer GlobalFoundries has been using for a while), so it could be better save us from another chip shortage (opens in new tab).
On top of all the industrial and scientific innovation, this year we witnessed what I believe was the first AI robot painter to hold an exhibition (opens in new tab) for the art she had created. And while she may not have been the first AI to enter the artistic field, she does provide a good picture of how artificial intelligence (even without a physical form) has made waves in the artistic community in 2022.
AI artist
Between Nvidia entering 10 million landscape images improved AI art tool (opens in new tab) in January so it could turn even my obscure scribbling into masterpieces, and Adobe decides this month that it will sell AI-generated stock photos (opens in new tab)it’s been a real roller coaster watching the AI art saga of 2022 unfold.
We’ve seen people use it AI to change Minecraft worlds (opens in new tab) in cute picturesque scenes, algorithms that extend works of art by interpreting what lies behind Michelangelo’s ‘Creation of Adam’ (opens in new tab) and our course came Meta who jumped on the bandwagon with his own AI video generator (opens in new tab).
Even our Dave trained Stable Diffusion to paint (opens in new tab) like his Uncle Hermann, which I must say is one of the healthier (and pretty cool) uses I’ve seen in the AI art space. Dave’s Uncle Hermann AI is actually worth mentioning, as it’s one of the seemingly rare instances where we can say with certainty that it doesn’t break copyright law.
The surrounding ethical discussion, exacerbated by events such as people use AI to win art contest (opens in new tab)s, has been quite heated. Many algorithms simply scrape the web and remix artists’ work without their permission, meaning thousands of copyrighted or otherwise private images are used without permission (as in the case of Google’s AI department being charged with using the health data of 1.6 million NHS patients (opens in new tab)).
Today it came to a head with protests erupt at Artstation (opens in new tab) over the phenomenon.
China has tried to address the problem by introducing regulations that tackle deepfakes (opens in new tab)Despite efforts to regulate it, we still seem to be in a wild west state. There are so many more unresolved problems (opens in new tab) what we need to talk about here.
Latent space
It was a weird time to explore what the techies call latent space.
One developer even discovered something very sinister. Loab would be the “first cryptid of the latent space (opens in new tab)In September, she emerged as a tired-looking middle-aged woman whose likeness seems to crop up again and again, even when the algorithm is presented with clues that weigh heavily against it. She often appears in grotesque scenes involving children and bloodshed, and lore begins to emerge around Loab as a child killer.
Was this algorithm fueled by those screwed-up chain emails people sent me as a tween? Who knows. One thing is certain, a lot of confused things are generated: the AI prediction of the “last selfie ever taken” (opens in new tab) being one of them. Although our collective pilgrimage into latent space can sometimes yield some truly heartwarming stories.
This month I spoke with Greg Linares, who uses a heavily modified local version of Stable Diffusion v1.4 to generate malware-based images, translate them into prompts, and feed them into the algorithm. It is the same malware behind the cyber attacks on Ukraine (opens in new tab), in fact known to most as the Sandworm group. A slightly unsettling inspiration, though I suppose it’s a way of dealing with something so ominous. He even told me that he had partially trained the algorithm that spat out these captivatingly dark images using photographs of his late father’s workshop tools.
First of all, good on him for using his own imagery, but it illustrates yet another fascinating way people are using AI image generation: as a sort of coping mechanism.
“It’s very cathartic,” he told me. And he’s not the only one using image generators to calm himself down. David Holz, founder of the Discord-based image generating app Midjourney, noticed a trend in generating images of dogs in the sky. It was clear from his conversations with the community that people had used the bot to process the trauma of losing a beloved pet. That’s the kind of tender request I can support.
Game design
In the same interview with David Holz, he prophesied that one day consoles will have a “giant AI chip and all games will be dreams.” (opens in new tab) A somewhat overzealous vision of the future that game designers are rightly terrified of, although some incredible leaps have already been made within the industry that allow game companies to work smarter and faster with the help of AI.
ZooBuilder, for example, is a prototype AI designed to animate animals (opens in new tab), to speed up the process of game development. And I talked to Scottie Fox who had been working on it AI generated worlds in virtual reality (opens in new tab). Maybe Holz wasn’t far off then.
We already have text generators and chatbots that are intelligent enough to work as one AI game master (opens in new tab), although these are admittedly quite resource intensive. AI Dungeon might just be one of those most demanding PC games (opens in new tab) in terms of GPU power we’ve ever come across, and for a long time it didn’t even have graphics. That puts into perspective how far we still have to go before we can at least run super-intelligent AI on our own desktop PCs.
And of course, generating text comes with its own moral turmoil. Kids are even turning to AI to write their homework. Although the same AI calculates Using AI to write your homework is a bad idea (opens in new tab).
feeling
Of course there is an entire corner of this minefield devoted to the problem of feeling.
I don’t have time to get into the nitty-gritty of what androids dream about, but I must give an honorable mention to Blake Lemoine: the Google engineer who thinks AI has become conscious (opens in new tab). Believe it or not, he has at least opened up some important dialogue. However, some rules have been put in place to prevent our overlords from getting too many rights. Since the US court ruled that AIs are not ‘individuals’ (opens in new tab) in any case, they are not allowed to patent anything themselves.
Of course, that’s what we’re most concerned about in a capitalist society, as opposed to the Blade Runner-esque moral implications such a statement could have in a synth-laden future.
Techno philosopher David Chalmers told me that, sure, AI NPCs in video games’can be conscious and can have feelings (opens in new tab)‘, and whether that’s true or ever will be true is likely to remain a huge bone of contention.
Either way, it would be a huge mistake to ignore the impact artificial intelligence has had on us as a species in 2022. While many of us may not be ready to fully embrace it yet, there is no doubt that its influence has shaped and will continue to shape the way we work, play and even heal.