Every now and then, someone tries to tell me how wonderful AI is. Obviously, I am not at all enthusiastic. I tell them that its praised abilities are problematic in every sense of the word.
The reply is usually that I do not know how to prompt it. I will not try to talk you out of using AI to generate images. But, as a crocheter, I want to explain how I have seen the pumping out of AI images compromise our perceptions of reality.
AI has proven to be a fantastic tool for crochet. I have seen it enhance images of crochet projects, visualise design ideas and suggest colour combinations. I understand why it can make the visual presentation of our work feel lighter.
Those who feed AI take cues from product photoshoots. Backgrounds, lighting and product positioning make an item look desirable; in the process, the product is given touch-ups so that it looks clean.
It is that attractive generated image that troubles me, because it augments reality and, many times, creates unfulfilled expectations for both the crocheter and the customer.
The AI image you generate would still need a crocheter to do all the work required to materialise the project. One has to select the yarn—its weight, type and colour—the hook and the stitches: technical choices reflected in crochet standards for yarn, hooks and gauge.
The time and energy a crocheter spends conceptualising an idea and choosing how to execute it are half of the creative process needed to complete the piece.
READ MORE: No Policy, No Shield: Why Tanzania Is Still Unprepared for the AI Era
Then they have to sit down for at least an hour and a half, even for a fast project such as a simple, solid-colour bucket hat. Crochet is a labour-demanding craft: when both hands are at it, no other physical task can be performed, and some complex stitches require uninterrupted counting.
As author Joanna Maciejewska recently said in an X post: “I want AI to do my laundry and dishes so that I can do art and writing, not for AI to do my art and writing so that I can do my laundry and dishes.”
Other activities may be skipped, rushed or postponed, especially for people balancing crochet with domestic duties. The point is the physical labour that is not as picture-perfect as the AI-generated image. With its ability to generate a multitude of images, AI presents a rosy view of a tedious process, to the point that customers may appreciate the value of handcrafting less.
What the image misses
Sometimes people have a hard time telling crochet, which uses one hook, apart from knitting, which usually uses two needles. I have seen AI-enhanced photos turn crochet stitches into something that looks like machine knitting.
New crocheters can then struggle to find the flattened stitches and assume that they belong to an advanced level because they look delicate.
Customers may start to believe that crochet stitches can be produced quickly and at the same price as factory copies. I have seen industrial designs copy crochet patterns, including the now-familiar imitation of the granny stitch.
But the feel and meaning behind handmade products remain; this may be why, even when baby clothes are readily available, parents and relatives still commission crocheters to make special ones.
READ MORE: A Genuine Lever or More of the Same? Tanzania’s AI Education Guidelines at a Crossroads
I have found that AI does not understand the relationship between yarn weight, hook size and stitches. It ends up making mishmashes of closely related but different crafts. Let me explain with the example of AI crochet wedding dresses.
To construct such a garment, you may crochet using double strands and switch hooks. The AI crochet wedding dresses I see, however, are based on fabric, which has different properties from yarn, such as drape.
To make a crochet wedding dress, one might combine yarns of different weights with fabrics such as lace, and use crochet flowers or beads to complete the bridal look.
A cumbersome exercise
I am not saying it is impossible to actualise an AI-generated wedding dress, but it would be cumbersome. Aside from the technicalities, the cost of making one would be through the roof. A handmade wedding dress involves a lot of trial and error, and even seasoned crocheters can take months to finish one.
As an heirloom, I would want it made from natural yarn, which is more expensive than the usual synthetic or blended yarn. Add the time and materials, then the cost of craftsmanship. Instead of reaching out to AI to give you ‘inspo’ for a project, maybe ask within the crochet community whether the project you want to create is feasible.
With the coming of social media, there are enclaves that have helped us see people with different hair, bodies, faces and colours. AI, I am afraid, is here to re-establish old beauty standards; UNESCO has found evidence that popular generative AI systems reproduce gender bias and stereotypes.
The difference between an AI-generated image that represents one standard of beauty and the way a garment actually sits on a body can be underwhelming.
The crocheter might think that something in their craftsmanship makes the garment less flattering. The customer could feel inferior compared with the beauty of an AI-generated model. AI images steal the joy of wanting something custom-made.
Human costs
We zoom in and out of AI images on our screens. Out there somewhere, AI has physical bodies. The infrastructure behind it requires electricity and has associated water and land footprints, as a United Nations University analysis has documented.
Even so, it is wrong to fault individual AI users for the environmental and human costs of systems they did not build. We use the internet, yet we do not advocate for ‘going off the grid’ because of the resources it depletes.
I think we should be self-aware of the consequences of our use but, more importantly, organise against the billionaire class that makes it possible for us to generate beautiful images while making Earth less habitable.
These AI images get in the way of creativity, connection and artistic production. In the meantime, maybe crocheters could put their phones down, make swatches, go to craft markets and sketch their ideas. The anxiety caused by wanting to create something as good as AI slop is unwarranted.
Dolly Parton’s Coat of Many Colors inspires me to create baby blankets. When I weave in the ends, I say a little prayer for the baby who will receive the blanket. In the song, Dolly says, “Perhaps this coat will bring you good luck and happiness.”
I have never commanded ChatGPT—or GPT, whatever the name is. If you know how to work it, I am asking for a favour: command it to generate a baby blanket inspired by Dolly Parton’s Coat of Many Colors. Ask it to direct you on the specific colours, yarn, stitches, hook and blanket size.
READ MORE: Tanzania Court Adopts Artificial Intelligence (AI) in its Processes
I am not sure whether the commands should be sent separately, because I also want to know what it will say about how long it would take to complete the blanket if you crochet full time, part time, according to the mood you are in, or while watching Korean series.
In monetary terms, can it calculate the price of the blanket, considering the price of the yarn, the complexity of the stitch, the skill level and the time? Here are my own blankets.
I learned from seasoned crocheters online and received help from other crocheters in Tanzania because AI could never…
Diana Kamara likes to introduce herself simply as the daughter of Adria Kokulengya. She is available at dianakkamara@gmail.com. The opinions expressed here are the writer’s own and do not necessarily reflect those of The Chanzo. If you are interested in publishing in this space, please contact our editors at editor@thechanzo.com.
