Close Menu
    AI offers a rare look inside the minds of CEOs—and can tell if they're depressed just based on how they sound on earnings calls
    Futurology

    AI offers a rare look inside the minds of CEOs—and can tell if they’re depressed just based on how they sound on earnings calls

    Share.
    View 18 Comments

    18 Comments

    1. Gari_305 on

      From the article

      >CEOs might be able to give away mental health challenges just by how they talk about their companies’ earnings to investors. A [study](https://onlinelibrary.wiley.com/doi/10.1111/1475-679X.12590?utm_campaign=publicity&utm_content=WRH_1_6_25&utm_medium=email&utm_source=publicity&utm_term=JOAR) published this month in the Journal of Accounting Research used artificial intelligence to analyze chief executives’ speech recordings to identify depression.

      >Researchers from [Indiana University’s Kelley School of Business](https://fortune.com/education/business/indiana-university-kelley/) and the [University of Kentucky](https://fortune.com/education/business/university-of-kentucky-gatton/) debuted a measure of identifying the severity and prevalence of depression among chief executives by training AI language models to analyze vocal patterns of CEOs. They analyzed data from more than 14,500 earnings calls from S&P 500 companies from 2010 to 2021.

    2. Baruch_S on

      Wow, we couldn’t come up with something better to use AI for than assessing the mental states of CEOs? Let them go cry on their yachts and wipe their tears with Benjamins. 

    3. prifecta on

      The human brain is still too complex for man made technology to decode accurately. Sure there are patterns and signs but too many variables like personality traits, psychological manipulation (some people portray to be more invested/concerned than they are to seem committed), health factors, personal distractions, etc

    4. Tekl on

      It would analyze homeless, but unfortunately, there’s no data because they are actually in the real world struggling.

    7. pat_the_catdad on

      I don’t need AI to know why I’m shorting Tesla with everything I’ve got due to watching Elon spiral out of control.

    9. blazelet on

      Something is driving their obsessive need to prove themselves and outperform their peers. Its not contentedness with life.

    10. Gold_Responsibility8 on

      Some people don’t even understand that this could be used to predict market moves

    12. tedfergeson on

      Two things that I don’t care about. Three actually. CEO’s, their depression, and AI.

    13. vonkraush1010 on

      Look at this point this is like those facial tic experts from 2 decades ago who claimed they could read people almost perfectly. It’s just impressive sounding hogwash.

    14. Unusual-Bench1000 on

      Good CEOs are actors of emotion. They can act up or down depending on the company weather. Real mental health is a private matter. You can really tell they were depressed when the window breaks and they fall out of it.

    15. artwarrior on

      You mean they are not happy after earning in one hour on Jan 1st what an average worker makes in a year? 

      Maybe their necks hurt from always looking over their shoulders?

    17. xFblthpx on

      I take it they have a lot of data of depressed people giving presentations?

      Dialogue involving presentation is very different in tone and cadence than phone calls or other forms of dialogue.

      I can’t imagine a model like this being remotely close to accurate unless such a data repo exists.

      Edit:
      Took a deeper dive and they do control for uninterrupted speech patterns but not specifically for presentations. Also weirdly they don’t account for time of day, which I imagine would have an affect on tone and intonation.

      The model they are using does have good accuracy on training data, but I’m skeptical of monologuing depressed people in private being close enough to the cadence of a very public earnings call presentation.

    18. Frustrateduser02 on

      I wonder what they had to sign or even were given the choice to opt in. Stock options maybe?