Researchers embracing ChatGPT are like turkeys voting for Christmas

Do you run your regressions by hand? Use tools where it makes sense to do so and where you get the maximum productivity gains. You don't need a PhD to get that, it is common sense. Where these models don't work, they often fail miserably. Much of the hype around this in academia are from people who have no idea what they are talking about.

"Do you run your regressions by hand?" You have missed the point of this commentary completely; probably intentionally so. It is not about the informed use of tools (digital or others). It is about the threat to our intellectual imagination. Lindebaum is rightly talking about the threats to the “intellectual craft” that is part of the process of thinking and understanding (including writing, which is not simply an activity of reporting results, but an integral part of thinking, creativity, and imagination so important for intellectual pursuits). An apt analogy is that of an artisan cabinetmaker honing her/his skills through practice while creating something new (and often novel and unique). Now, you can automate the production of cabinets, but the artisan skills are lost along the way (and the entire makeup of a cabinet is adjusted to the fit requirements of the machines and processes set up for producing them). The artisan craftsperson is turned into a worker simply operating them. This transformation involves substantial deskilling (of the craft) and reskilling (to operate machines); nonetheless, a substantial change. Workers are basically reduced to being mere operators (and if you further automate, entirely replaced by autonomous robots sooner or later). In academia, we will see more “unthinking” reporting of “structured reviews” of this or that, shallow research results of spurious causal relations and variables tested with the help of statistical tools; statistically sound and methodologically sophisticated but conceptually uninspiring and often suspect. Most of the stuff published in academic journals these days is intellectually underwhelming, lacking real novelty and true imagination (I can only talk about my own discipline of business and management, though). The use of AI and tools such as ChatGPT will further accelerate this process (and it will come, I have no doubt about that). It will greatly advance the careers of the “career minded” academic, who are already focusing on pleasing metrics, hitting productivity targets, the quantity of output and the purported “sophistication” of the tools used, pumping out one SEM-based nondescript study after another, for example. They are already akin to being academic “factory workers” rather than “intellectual craftspeople”. If the above get eventually automated away by AI tools, so be it. Maybe you are right, and it is not a bad thing in the end. The deluge of outputs (due to the publish or perish game) hardly gets read anymore already. If we eventually automate and outsource the entire process of academic writing, reading, and teaching, we could be so much more efficient, pumping out even more stuff that nobody reads. The AI can do the lecturing and conference talk too. We will be able to send our AI avatars to such events. Much more efficient this way, who has time for such ancient nonsense. Eventually, AI will talk to AI. Finally, we will be free to do things that really count. Unfortunately, we will have lost the essential skills required to do something intellectually meaningful by then, along with the understanding of why such intellectual skills are relevant in the first place. Be sure, the rich and powerful will continue to have access to intellectual craftwork (and skills), but it will be intellectual fast-food for the masses most likely.