While the study looks fine from quality standpoint, I feel like the whole idea was stacked in the “AI” favor. They chose the task of generating ideas. Which is one of a few things that LLMs are good at as the name implies. They can spit out unlimited amount of ideas for a person to choose from, while without it people use something like a word cloud which is a lot more time-consuming on the front end, but leads to better ideas from the start.

The whole “social connectivity” aspect is too subjective to offer any conclusions.

Treat LLMs as a tool, like you do internet, and it can be useful, but it will never be more.

Full study: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5188231