Would humans be replaced by AI in data science?
I have seen people so anxious to detect AI on writing that they started to use simplistic and useless technique. “as of my last knowledge update” is said to belong to AI-written text. What about a human using this sentence? I would use. Would it be worth raise accusation on an author? Our paper is serious for us. It seems a desperation towards an AI each time smarter and smarter, harder and harder to separate from humans.
No. there is no evidence that human experts will be replaced; what will most likely happen is an automation of basic tasks, that are already automated by libraries. Coder Interpreter is using Python libraries, what it is doing is deciding which tools to use, given a rich menu, and using its language capabilities as LLMs to interpret the results, but it does not replace humans as final end. What I see is a new way to make data science, something that already happened when those libraries used started to gain momentum. It is just a step forward towards something already in course: making repetitive, but important, research steps into straightforward tasks, without losing scientific rigors. Most data scientists are not necessarily computers and mathematical sorcerers. Thus, the fact that they can perform important statistical analysis without needing a multidisciplinary team, or even, needing to know the basics all the time is something to celebrate. Of course, we still need human’s expertise and guidance, deciding what to prioritize. Even, in some cases, the correct conclusions. Human will not be replaced, it will just have more energy for asking the big questions. Coder interpreter is a productive tool, not a artificial general intelligence.
