💡 Position paper on the danger of deskilling through AI
When discussing AI and skills, the focus is often on upskilling or retraining. A topic that is not so often the focus in the discourse around AI education is deskilling. Deskilling is the phenomenon that with AI systems, certain skills are no longer used as often and atrophy. A typical example is text translation, which is now so good with AI systems that humans only need to check the translated text for typos or stylistic problems. Or in ChatGPT, idea generation is often taken over by the system and humans have much less incentive to come up with original ideas themselves. To draw attention to this problem, Professor Gabi Reinmann from the University of Hamburg has written a position paper about the possible disadvantages of using AI. If you want to get a deeper insight into deskilling, I recommend the article by Advaid Bhat and colleagues, which explains how Large Language Models influence the writing process. In my opinion, there is a fine line between AI deskilling and complementing existing skills. Skills are often represented as a tree structure, with complex skills at the top. Take writing a book as an example: this high-level skill is unlikely to be replaced any time soon and is not at risk of deskilling, but any other subskill, such as coming up with an idea for a chapter, could be at risk of de-skilling. Nevertheless, I would argue that de-skilling is often accompanied by a newly learned skill that supports the high-level skill, e.g. getting good ideas from large language models by prompting them.
🔗 https://lnkd.in/gkRvRD5C
🔗 Advaid Bhat and colleagues: https://lnkd.in/gGeiEs4e
#aieducation appliedAI Institute for Europe gGmbH
"AI will make our lives better" :) ich glaube fest an viele Chancen, auch wenn es Herausforderungen geben wird, gerade aus der People Management-Perspektive heraus Nina Kataeva Dr. Henning Schierholz Christoph Helm Dr. Martin Stotter