Skin In The Game

Quite possibly my favorite part of Blaugust this year has been Tipa's fascinating series on AI generated content. Tipa asked "OpenAI’s GPT-3 to generate 31 different videogame ideas, one for each day of Blaugust, each dealing in some way with that particular day of the month", a process which created some amazingly convincing games I'd definitely play as well as some extremely bizarre ones I certainly wouldn't.

Even more intriguing than the games themselves are Tipa's sidebars and posts about her methodology and how much prompting, tweaking and editing it took to get to the finished results. On the evidence presented, I think it's clear that publically-available AI text generators aren't going to be putting many human game devs out of business any time soon, although they might come up with a few game ideas we haven't seen done to death already.

I find all of this incredibly interesting, possibly due to growing up reading too much Philip K. Dick. I always wanted to live in a world with talking taxicabs and toasters that argued back. 

Now that we very nearly do, I find myself somewhat ambivalent about it all. I have an irrational dislike of AI assistants like Siri and Alexa that's borne not out of any kind of fear of what they might be doing with my data but of a deep sense of disappointment that they're not better. I want them to be more like the AIs in Iain Banks' Culture novels, less like a flock of robotic Captain Flints.

I've already played around some with AI-generated images and text . I'd been using Craiyon for pictures and Inferkit for words, the big attraction being that they're both free and require no regestration but Tipa's posts have inspired me to dig deeper.

Tipa has been using GPT-3 for writing and Dall-E 2 for pictures. They're both operated by OpenAI and are paid-for services, although as Tipa explains, the fees are very reasonable and you get some free use as well. 

You can register to use the text services right away. I signed up yesterday and played around with some prompts this morning. I'd love to start experimenting with Dall-E 2 as well. From the results I've seen it's hugely superior to Craiyon (Hardly surprising, given Craiyon was originally called Dall-E Mini.) Unfortunately, access is by invitation only. You can ask for an invite but there's a waitlist. I filed out my application form almost a month ago but as yet I've heard nothing.

Before I started messing around with GPT-3 today I logged into Guild Wars 2 for the first time in over a month. I wanted to claim my free mount skin and all the other not very exciting goodies ArenaNet is handing out for the tenth anniversary.

I'm very much on record as not liking GW2's mounts but I'm even more on record as not being able to resist a freebie. I'm not currently playing World of Warcraft either, nor likely to any time soon, but that didn't stop me re-installing BattleNet just to claim my free copy of Shadowlands and my Level 50 Boost.

After I'd patched GW2 I logged in and looked through all the options. There were a handful I quite liked but I decided to hang on to my free skin for now. I'll almost certainly take one of the big cat options for the Warclaw, the World vs World mount, since that's both the least obnoxious to ride and the one I use most. 

Next, I logged into OpenAI, and after I'd had enough of getting GPT-3 to tell me stories about dogs and cats on Mars, I thought I'd see what it had to say about ANet's underwhelming anniversary bash. The prompt I used was 

Write a blog post describing the free mount skins being given away in the mmorpg Guild Wars 2 for the 10th anniversary. There are skins for all the mounts. Focus on the skins for the warclaw mounts used in world versus world.

The result was admirably on point:

"In celebration of Guild Wars 2's 10th anniversary, ArenaNet is giving away free mount skins to all players! This includes skins for the new warclaw mounts, which are used in World versus World.

The warclaw skins come in four different colors: red, blue, green, and yellow. Each one is uniquely patterned and sure to make your mount stand out on the battlefield. Whether you're leading the charge into battle or sneaking behind enemy lines, these skins will make you look good doing it.

So log in to Guild Wars 2 and claim your free mount skins today! And don't forget to check out the other 10th anniversary celebrations going on throughout the game."

Apart from the "four different colors" bit, that could almost come straight from an official press release. It's very similar to the emails games companies send me regularly. If I was a busy copywriter I could very well imagine saving myself a ton of time using this thing. Or I could imagine it replacing me entirely.

For comparison, here's the result from InferKit using the same prompt:

"Accounts with two or more followers are eligible to win.

The contest is open until 2PM PDT on August 5th. You must be able to submit and post the contest in the forums to be considered.

There will be four winners selected based on the quality and creativity of the blog post, creativity in responding to the prompt, quality and creativity of the blog post description, and the amount of followers a follower has.

The winners of each Skin Prizes:

The first prize winner gets a Mecha Troll Mount"

That's much weirder. And utterly wrong. I do want a Mecha Troll Mount, though.

Finally, I asked Craiyon to give me some pictures. My prompt was fairly vague:

Guild Wars 2 Warclaw mounts in various different "skins".

 
Based on my previous experiences with Craiyon, I wasn't expecting much but the results were surprisingly good. Leave aside the distortion and most of those could be quite decent skins. What's more, at least two of the shots clearly show some distinctive hot keys along the bottom, just where they'd be if you were playing, which made me think they must be ripped from actual GW2 screenshots.

I ran a second prompt, replacing "Warclaw" with "Raptor" and this time it was undeniable. Several of the shots are just blatant steals from in-game screenshots, complete with UI. Oddly, the raptors themselves look considerably less like anything you'd see in game.

That raises the awkward question of attribution. And ownership. There's increasing concern in some quarters about the way AI programs collect and repurpose data they find on the web. The process is opaque and the results are of unknown provenance. Artists and writers were already concerned about the potential for AI creators to steal their commissions. Now they're worried they may be stealing their work as well.

It shouldn't just be the creators who are worried. People like Tipa and me, playing around with the tools and sharing them online, might well find ourselves caught up in any copyright or IP rights infringement arising from the dubious ethics of the machine minds. I hope the good old "fair use" excuse still holds.

It all reminds me enormously of the early days of sampling back in the 80s and 90s. For a while there it was the Wild West, everyone grabbing whatever they fancied and using it how they liked without a thought to where it came from. 

All new tech starts like that but lawyers and legislators are getting faster at closing it down. The time to go wild with autonomic mashups may already be coming to an end. 

I hope not. I feel like I'm just getting started. I also think the fears are greatly exaggerated. Creativity isn't a zero sum game. We can always use more. And every creator is unique. Name your favorite artist or writer. There's never going to be another quite like them. 

Also, the likenesses aren't always very good. The prompts for the laughing and flying mounts specified "in the style of " named artists. I defy anyone to tell me who they were. And what do you mean, you didn't notice they were laughing?

On the other hand, as my first example above suggests, if you're looking to leverage that liberal arts degree into an entry-level job in video game promotion and copywriting, maybe think again. That ship might just have sailed.

Art For Art's Fake

Late last night, Wilhelm alerted me to an internet fad that's apparently been all over Twitter for a couple of weeks. Not that I'd know. I'm not one of the Twitterati. 

Maybe I should have known, even so. When I followed the links, first to the version of Dall-E that Wilhelm favors and thence to Craiyon, its new home, I realised the whole thing was part of something I already have bookmarked, the GPT-3 Demo Showcase.

As with a lot of aggregators, there's just too much to wade through there. Over three hundred AI-related apps using the GPT-3 large language module made publicly available by Open AI last year. I'd been riffling through a few of them in idle moments, playing around with various "writing assistants" and chatbots, but I'd been under the impression everything there was purely textual.

It seems not. Craiyon is an "AI model that can draw images from any text prompt", something like the Artflow app we all had so much fun with a few months back. Except that where Artflow only generates portraits, Craiyon attempts to visualize anything you suggest. 

It's not content with a single try, either. Craiyon gives you nine different versions in a classic comic-book grid, something which creates, in my mind at least, an almost subliminal expectation of some kind of narrative progression. It adds to the general surrealism of the process, as if it wasn't surreal enough already.

Even though it was late, I couldn't resist playing about with the thing right way. I did a couple of runs at the Dall-E Mini site before moving to Craiyon, beginning with this one:

 

I wasn't entirely convinced. Most of the shots look a lot like Tony the Tiger but not much like they've been drawn by Bill Watterson, the creator of Calvin and Hobbes. #7 looks something like Hobbes. I'm not sure any of the others do.

I was curious as to why several of the shots have distorted, almost cubist faces. Nothing in the seed should have suggested it. As things progressed it became obvious at least a degree of facial distortion was just about guaranteed, no matter what seed was used. I still don't know why. Maybe the AI's on something.

Still on the tiger trail, I tried Shere Khan next. Once again, I couldn't really see much - or any - of Bill Watterson's influence in the result. I started to wonder whether Craiyon really knew who Bill Watterson was or what his drawings looked like. Of that, more later.

There followed a very unsuccessful series of attempts featuring Sabrina the Teenage Witch and Gidget. In only one was either of them even vaguely recognizeable. This one.

 

The AI seems to think "various" means "identical" but it's very keen to prove it knows what a witch is. In # 4 it seems to have given Sabrina two hats to wear and by the last frame it has at least managed to come up with a color other than black. Still, not very impressive.

 

Back to the drawing board, by which I mean more cartoon characters. I was wearing my Stience T-shirt so Bean from Disenchantment was the first one I thought of.

It seems that being the lead character in a successful Matt Groening animated series is no more a passport to fame than landing a long-running role as the face of a once-feted mmorpg. The AI had no more idea who Bean was than it recognized Firiona Vie's name when Wilhelm pointed it in that direction. Of course, the fact that I got the name of the show wrong (As I nearly always do.) might have had something to do with it. 

 

[Edit: I just reseeded it with the correct name and got this:

My apologies to the AI. I am an idiot.]

It clearly knew what a  kite was, anyway, so I tried again with someone I thought was bound to be famous enough to be pictured flying one: Elvis. 

I really like those. It's nice to see Elvis looking so happy. 

 

Since the AI was clearly at home with Elvis, I thought I'd give it a chance to show off. It was plain from Wilhelm's post that one of the app's great strengths was cloning the signature styles of famous artists. If Bill Watterson wasn't famous enough, how about Banksy? The internet loves Banksy.

Wow! That's just about perfect. It would make a great poster. Or a mural.

It was at this point that I had a brainwave. Maybe Craiyon didn't know who Bill Watterson was but how about his most famous creations? How would it be if I asked it to draw Elvis not in the style of Bill Watterson but in the style of Calvin and Hobbes?

Oh boy. It loves Calvin and Hobbes! So much so that it barely remembered to put Elvis in at all. That's not "Elvis in the style of Calvin and Hobbes" so much as Calvin and Hobbes with a very brief cameo by the King.

Since Craiyon was so into them, I gave it another chance to indulge itself.

 

 

Apart from the distortion, those are very good fakes. It wouldn't be all that hard to add some speech bubbles and turn them into a strip. It's easy to see the potential for mischief, something Calvin might love but I'm pretty sure Bill Watterson wouldn't.

 

I thought maybe it was time to move away from cartoon characters and copyrighted material. Since Elvis worked so well, how about a couple more rock stars?

 

I didn't specify "...during a pandemic" but Craiyon decided to add masks anyway. Other than that it's a mind-blowingly accurate visual record of an incident that never happened and probably never could have happened. I'm particularly impressed by the way it correctly interprets the preposition in all but two of the pictures.

Time to step down hard on the surreality pedal. I'd been keeping my texts short and to the point. I decided to go long and complicated instead.

Once again, I was impressed. "Ann of Cleves" doesn't feature but then I did spell her name wrong. I guess you can't tell it's a Thursday, either, or what they're talking about, but it's definitely night-time, they're obviously having a discussion and that could as well be The Factory as anyplace else. The rabbits, the monkey, Nico and Andy are all clearly recognisable, even if facial distortion is at an all-time high.

It freaked me out a bit. It's like set of contact negatives from a photo-shoot for Interview magazine, one that was judged too disturbing to publish. It was getting properly late by then so I decided to call it a night before things got any weirder.

This afternoon, just before I started on this post, I ran a couple more ideas past Craiyon. Rather than leave you with the unsettling images above, here's something decidedly more cheery. And colorful. Boy, is it colorful!

 

Doesn't that have a weird Rembrandt/Vermeer feel to it? Ferris looks as if he's wearing a ruff in #2 and a cassock in #7. There's an awful lot of Scooby and not so much of the gang. I can make out Daphne in #3 and Velma in #5, I think, but if Fred or Shaggy are in there anywhere, I'm afraid I can't see them.

 

Saved the best for last there. I'd have that on my wall, blown up to poster size. That and the Banksy. I love the way Scooby seems to be wearing clothes in most of his shots.

As I said about Artflow, this would really integrate wonderfully with Occupy White Walls. I really ought to look into that. I think it might be doable. Whether or not it would be legal is another matter altogether.