Discussion
I'm an absolutist, getting a lot of value out of not using AI for anything, and so didn't connect with the requisite 'it's a balance' slant in this article. Nonetheless the research itself looking at usage trends is very interesting indeed.
https://hbr.org/2026/06/how-people-are-really-using-ai-in-2026
@JulianOliver I'm rather fond of usage #9 'Astrology and tarot readings'. There has yet to be discovered a better use case for the generalising, aproximating, and hallucinating text machine.
"You're lucky number today is 3, Julian."
@JulianOliver
"50,000 records that were collected between March ‘25 and Feb ‘26 from a wider pool of online sources, adding LinkedIn, TikTok, and YouTube to our previous sources of Reddit, Quora, and articles. "
I don't think this enables real insights as these are only people who decide to write about their experience and therefore it's very likely very biased and self selected
@realn2s The authors do confess the challenges of working with information collected from these sources. I find it interesting nonetheless, a pool broad enough to give some insight into usage trends.
@JulianOliver
I saw it but i couldn't read it as it's behind a paywall. It's good that the authors are aware of it but IMHO doesn't change that fact that it doesn't live up to the headline
And I'm also aware that any other kind of study is very hard without the cooperation of the ai firms (which would possibly taint the results in other ways)
@realn2s The samples would be prone to no less bias were they to solicit statements of usage, as a poll or interviews, from the general public - esp on this topic. Also done this way it may be even more prone to pollution from bots.
Working with AI firms aside, they would get a flatter pool if they were to work with organisations and/or business that impose engagement across their base. But that can be a big ask for a small research project.
Install bonfire.cafe
Get the full app experience
