ML workflow using LLMs

am just looking for comment/feedback here. So, from one of Andrew NG’s courses i gathered i can instruct an LLM to run a machine learning algorithm…
So i prepared a dataset fit for a Classification problem and gave it to Bard, instructed Bard to do some cleanup and feature engineering, and run an ensemble. But when Bard finished all, it did not want to share the results because it thought i would take financial advice from there … odd… so i just got a response that said the 'model performed generally well’l but i cannot share results.
Anyone has tried the same? or is it not a good usecase for LLM?

Bard does not want to be legally responsible if you use its results to make an investment and then lose all your money.

Thanks Tom. I tried to persuade Bard that my goal was a research project but did not work out. I am guessing what gave it away is that i mention it’s financial data i was working on… i’ll revise my prompts…