ChatGPT: If Statistically Significant was good enough, why ML/AI took over Statistics?

\begin{PPASTA}

ChatGPT and other similar chatGPT-like tools and models have rapidly shown their benefits and applications, mainly to generate sounding-right, statistically-significant word sequences based on their data oracle and the context (the user's input at the prompt). Interestingly, Statistical tools also behave in a similar manner.

Now back to the question: If Statistically Significant was good enough, why ML/AI took over Statistics?

Statistical tools are powerful with their own application. Especially for those businesses that can afford a multi-decision making operation, these models and tools would allow the business to explore the decision making space while keeping the risk at a level low. In contrast, small businesses, who might have only one shot, would rely on ML/AI models to hit a niche accepting the risk. Please note that here we figuratively differentiate Statistics and ML/AI.?

This note is not to question the capabilities and potential of chatGPT-like tools (which are still rapidly evolving, especially via combination with other tools). Instead, it is to express a dislike to see a problem statement is being rewritten just to make these presently-available, amazing tools its solution. Having public speaking skill is an advantage for any mayoral candidate, but it should not become the sole requirement. This might also be how Arts and Decision Making separate from each other.

Just a thought.

\end{PPASTA}