Are AI Images Widely Accepted in Malaysia Yet? Here’s Some Case Studies
TL;DR
- For fun: OK
- Commercial but artistic-looking: OK
- Commercial but imitating human drawings: Nope
- [BONUS] Commercial but lame: seemed OK?
The Cat Marathon
In Mid-May 2023, a post about an alleged “cat marathon in Kuala Lumpur” went viral. The pictures were posted in a MidJourney group, so it was without intention to hide the fact that they were AI generated.
The outcome? The public loved them! So much so that everyone wished it was a real event!
AI-Generated Covers: A Positive Case
Late May, Buku Fixi, a local publishing house famous for their fictions, published a series of hard-cover limited edition of their best-selling horror novels to be sold in the Kuala Lumpur International Book Fair. They were AI generated too, and they were open about it.
The reception seemed to be pretty OK as they were all sold out as far as I know. No rage detected on Buku Fixi Facebook page. In fact, they have even tried the same in May 2022 using Wombo AI, two months before MidJourney even went beta.
With these positive cases, it’s easy to think Malaysians are comfortable towards AI images. However, all that perception changed soon after.
The Charred Burger
Just a few days later, myBurgerLab, a burger chain known for their charcoal burger buns, posted an image captioned “if myBurgerLab was an anime”. It seemed that the chain owner was all excited about the image outcome, but contrary to his expectation, it flopped.
Negative comments protesting the use of AI image generators especially from the art community poured in like a waterfall. The owner took note and apologised.
The Lame Comics
On the same day, AEON Malaysia, a retail chain from Japan, posted a series of AI generated comics too. The comments were negative as well, but not entirely about the usage of AI.
Why? Because the story didn’t make sense!
It was really funny that the lame storyline diverted everyone’s attention and people kinda stopped complaining how it’s AI generated.
AI-Generated Covers: A Negative Case
Just recently, a local publishing house called Kadokawa Gempak Starz published a few books under their imprint “Magic Bean”. They were alleged to feature AI generated covers but went under radar. Nobody complained.
While it seemed OK, just last week, Magic Bean announced that they have published the first ever AI-generated picture book in Malaysia, titled “The Butterfly Catcher”.
The first few days were quiet, even anticipated…
A few days later, a group of people started the rage, presumably from the illustrator circle. Well, Kadokawa Gempak Starz is famous for their comic books after all, so naturally it is the top career choice for many local comic artists, and this upset them. The protest grew, and the announcement post was quickly filled with negative comments.
The community even started to point out that Magic Bean’s recent publications (mentioned above) seemed to be AI-generated. It was so fiery that the publisher decided to pull the plug and take that book off the shelves.
The Book in Question
Incidentally, I got a hold of the book.
Knowing the author (not personally but he’s rather famous), he does have a strong foundation in art and he used to draw web comics sometimes.
If you are used to observe AI images, you could see that the images in this book, other than a few occasions of less-than-normal finger count, don’t actually feature a lot of the usual AI anomalies.
The picture book in question has consistent character design, matching water reflections and some recognisable hand-written words. These traits are almost impossible to achieve directly in the current state of AI generators without human edit.
I’m not saying that using AI is good, I’m just saying it’s obvious that the author must have spent some time and effort to fix those anomalies and to add some human touch in them.
My 2-Cents
From business point of view, it’s very very tempting to use AI image generators as it could save not only money but also a lot of communication time. But we all know that most of the AI models today are trained on images salvaged from the internet without the creators’ permissions.
Will my company (also a publishing house) use it despite the rage against it? Let’s be honest, we probably will, to a certain extend. It’s such a struggle between ‘we could’ and ‘we should’.
