It’s WIP Wednesday and this week’s theme for #showmeyourknits is cardigans! Fortunately my wip this week is the Pivoine Cardigan by Audrey Borrego, a drop-shoulder cardi with lovely lace detailing at the bottom of the torso.
Borrego recently stepped away from the knitwear design game a few months ago, and very generously made all her patterns free on Ravelry. When I checked Ravelry that day and saw that she was all over the first three pages of the Hot Right Now feed, I had the same thought as some other folks: “Holy shit, did something happen to her?” Nope! Just a voluntary change in life direction. Hero portfolio of 100+ patterns is an amazing gift from a very prolific designer, and I hope she’s having a wonderful time in the next chapter of her life.
I’m using Knit Picks Stroll Tweed, which I picked up a few years ago during a Halloween sale (they discounted all their black, orange, green, and purple yarn). The pattern calls for sport-weight, but considering the suggested needle size for the body is US 3, a fingering like this is working nicely as a substitute. I’m knitting size 7, and the depth of the armholes is perfect – I really like wearing big t-shirts, and there’s plenty of room for my tee sleeves to fit without weird bunching at the shoulders.
I also plan to add significant length to the body so it covers my butt, and now that I’ve passed my natural waist I’m adding increases to accommodate my hips circumference. I’m adding increases at the same rate as the pattern calls for decreases in the sleeve, and I’ll just have to make sure I stop when my stitch count reaches the correct multiple for the lace pattern.
There’s a possibility I’ll run out of the yarn before I get to the sleeves, but honestly, I’m hoping that will be the case. I have plenty of other fingering-weight skeins in my stash that need to be used, and I think contrast sleeves would be very striking for this project. The stash bust continues!
And now, a lightly cuss-seasoned rant about generative AI
While on Ravelry this week I came across a “design” whose picture was clearly generative AI and reported it. As far as I can tell, it’s been taken down. I figured Ravelry wouldn’t be AI-friendly (I was not the only person to report the page) but it would be great if the mods could make a policy statement about AI use, if only to encourage its users to take a broader critical stance against generative AI.
Obviously AI in knitting patterns is a problem. If the “designer” can’t even produce a finished object from their own written pattern, they’re clearly a scammer and unable to write a pattern to begin with. There’s no reason to trust that the pattern will be any more real than the generated image they’re using to advertise. But generative AI is a much larger problem that has infiltrated pretty much everything we interact with on a daily basis. I believe it poses extremely bleak outcomes for creativity and critical thinking.
Before I go off I want to point out that I’m not immune to the draw of AI and how it can be used. When I was studying creative writing, I wrote an experimental poem using the predictive text AI on my phone. I presented it to my workshop, completely transparent that I had used predictive text, because I was intrigued that the data my phone had stored about my typing history created a digital mirror of my personality. I called it “Runecasting with Predictive Text”. This was back in 2017, when generative AI was still producing eye-burning fractal nonsense images that could never be confused with a photograph or human-made artwork. Now I look back at that little experiment and shudder. I want to share my shame with you.
Generative AI is way more refined now. It’s replacing human labor, especially creative jobs like writing, art, and design. It’s widely used in political campaigns to smear opponents and create propaganda. It’s even being used to write public policy that we’re all supposed to live under. NYC Mayoral Candidate and sexual predator Andrew Cuomo faced criticism for using ChatGPt to write parts of his campaign’s housing policy. And yet our social media networks, job hunting sites, even this blogging platform I use proudly advertise how they’re “harnessing the exciting potential of AI” or some such bullshit to appear cutting edge to its users. The little sparkly AI icon is menacingly ubiquitous.
Left: WordPress invites me to generate an image with AI instead of using this screenshot. Right: The aforementioned sparkle icon.
AI provides no new information to us. It plagiarizes artists and writers. It only reflects, and all it reflects is existing human thought, including our worst prejudices and bigotries; it just delivers them back to us faster. And if it’s used by elected officials that are too fucking smooth-brained and lazy to come up with new ideas or to think critically, then I don’t want to live here anymore. And we probably won’t live here for long anyway, because generative AI is pumping massive amounts of CO2 into the atmosphere and sucking up our water supply.
Last, I want to address the argument that generative AI helps people be creative. Nothing could be less true. Artists love the process. They love practice. That’s why great masters’ sketches are valued by art collectors and museums, not just their finished works. That’s why there are so many musical compositions called “etudes” (studies). It’s why dancers and athletes and musicians and anyone who uses their body to do anything has to spend so much time conditioning their muscles and joints for performance. If you have an idea for an image and want it to exist, and you use AI to make it exist immediately, then that isn’t artistry – it’s spectatorship.
There’s nothing wrong with being a spectator. But the ethical thing to do is hire an artist. Support artists and writers who are using their brains and bodies, who are making truly creative and critical decisions about art. Or do it yourself! Actually do it yourself. Accept that your first try at a creative endeavor might be messy, poorly executed, or embarrassing. No one needs to be perfect. We just need each other to try.
If you’re newer to the concept of generative AI or just want some help spotting it, here are things I look out for.
In images
- Look at any letters, fingers, or (in the case of knitting and crochet patterns) stitches. AI has a harder time generating fine details like these; words will often be illegible, hands will have too many/too few fingers, and stitches will not make any physical sense.
- The depth of focus is really shallow. Gen AI will often blur backgrounds and even the edges of a subject to reduce the amount of detail it has to create.
- Everything is oddly smooth. If the image looks like it’s lacking texture and human figures look like they have silicone for skin, it might be AI.
In videos – many of the above red flags apply
- Multiple people seem to have the same face or look extremely similar.
- People moving toward the “camera” do not appear to be getting any closer.
- A person’s face is moving but their body is not, or the head and body are moving out of sync with each other.
- A celebrity starts talking like they’re an SEO-optimized ad – they start talking using words like “viral”, “Hollywood”, “diet” in relation to a specific product. It’s likely a deepfake using actual footage of the celebrity with a generated voiceover. Obviously anybody could actually be saying these words, but if it’s in relation to a weight loss plan or beauty product, exercise caution.
In writing
- The author’s portrait appears AI-generated. Unfortunately publications are faking writers along with plagiarized writing.
- The same ideas are repeated in re-worded phrases or sentences.
- The sentence structure is repetitive.
- There are a lot of words over-used by AI; this Reddit thread has a good compilation of examples.
- The website or publisher has recently laid off large portions of its staff; it’s likely they’ll rely more on AI now that they’ve cut their human workforce.
I think that’s all for this week. Until the next WIP Wednesday, my dudes, and watch out for that AI slop!
Smitty is Knitting 