[OC] Arbeitsplatzbekämpfung mit KI -Bildgenerierung
Von datacommissar
![[OC] Arbeitsplatzbekämpfung mit KI -Bildgenerierung](https://www.bytesde.com/wp-content/uploads/2025/08/7514p7fuf1hf1-844x1024.png "[OC] Arbeitsplatzbekämpfung mit KI -Bildgenerierung")
[OC] Arbeitsplatzbekämpfung mit KI -Bildgenerierung
Von datacommissar
10 Kommentare
Data source: Felten, E., Raj, M., & Seamans, R. (2023). Occupational heterogeneity in exposure to generative AI
Tools : Created in R and ggplot2 edited in Illustrator
So, if your job involves making images, art, or designs on a computer, AI is coming for your work first.
Artists, graphic designers, interior designers, and architects; AI tools are already doing stuff you used to get paid for. The more you rely on pictures or visual creativity, the more you need to worry about losing work or having your job changed by AI. But here’s the kicker, these AI tools replacing jobs aren’t free of cost, they are incredibly expensive
Cool! I’m just confused why some of the extremes don’t have a job title referenced? In Computer & Mathematical we have Support Specialist listed but no title listed that is least impacted.
What is the y axis?
Why there is a random arrow in the „Education, Training, and Library“?
This data seems to show „exposure“, but on the title this exposure is claimed to imply „reliance“. This is IMO a wrong conclusion, exemplified by mathematicians being one of the most exposed, and I would strongly claim, one of the least reliant on AI generated images.
I like how you’ve considered make it look good, but there’s not enough contrast between the yellow and green, and the overall color choice is terrible for colorblindness.
what is the central gray line?
what criteria are used to highlight the red subset? it better be good, otherwise it appears to be cherry picked
I’m pretty sure ”Mathematician” is high on the list because most image generation is *not* actually important, so can be safely replaced with a random AI image, not because mathematicians are highly reliant on AI, as would be implied by the description. You would still use LaTeX for figures that actually have mathematical significance to them, since in those, a single error can undermine the entire argument.
I’m pretty sure that pure mathematics would be one of the hardest topics for AI to master that relies solely on mental processes rather than physical tasks; it may one day, but certainly not today.
The selection of red highlights seems arbitrary
This seems very skewed and I do not necessarily find the method of calculating such „scores“ very helpful, I might even go so far as to call the method used to obtain this data a bit arbitrary.
You state „…that AI can potentially automate or augment“ the potraied tasks, but I don’t necessarily agree that this is what the publication is stating
Overall, it seems you went ahead, showed some data, made some „uh look, AI, scary“ clickbait title but failed to explain the data
And the data itself is from a publication of an author that cites his own method to calculate the data and frankly that method seems a bit yeah… Arbitrary
Overall, not a fan of the data/message but pretty pictures, I will give you that
Are you using the data from the table in the supplementary materials 2? From my quick scan, those results are exposure to all generative AI, not image generation. I am actually surprised they propose that there is such a high correlation between professions susceptible to all gen AI vs. just image generative AI.