9 votes

Instagram photos reveal predictive markers of depression

9 comments

  1. [8]
    NubWizard
    Link
    Very cool technology. It seems like a neat proof of concept. So it seems like we can potentially use these techniques to flag risks of depression but the questions I guess is should we and who...

    Very cool technology. It seems like a neat proof of concept.

    So it seems like we can potentially use these techniques to flag risks of depression but the questions I guess is should we and who should?

    I need to do a deeper reading into the article to see if those questions were asked by the researchers. I don't think I would want to give Facebook the right to take implicit action on my mental health. I could see it being a useful tool though for therapists and mental health specialists as a way to get an understanding of when depression begins and have a conversation about the events surrounding the timeline.

    But I would hate to see the capitalistic outcomes from this research.

    6 votes
    1. [3]
      EscReality
      Link Parent
      It would be an interesting thing to implement as an automated system on a photograph based website like Instagram. Have a bot send a message if an account meets certain criteria for x amount of...

      Facebook the right to take implicit action on my mental health.

      It would be an interesting thing to implement as an automated system on a photograph based website like Instagram.

      Have a bot send a message if an account meets certain criteria for x amount of time.

      Something along the lines of;

      Hey 'user' I just wanted to check in and see how you are doing, if you need to talk to someone we are available 'here'. If you need anything please let us know.

      Keep it subtle and kind.

      2 votes
      1. [2]
        pleure
        Link Parent
        I don't think this is a good idea, at least personally I want people I know to talk to me. I appreciate the sentiment beyond strangers who try to help and have spent quite a lot of time...

        I don't think this is a good idea, at least personally I want people I know to talk to me. I appreciate the sentiment beyond strangers who try to help and have spent quite a lot of time "counseling" depressed strangers online myself, but ultimately I think help needs to come from friends and family, not strangers, and especially not robots.

        1 vote
        1. EscReality
          Link Parent
          Yea I don't really think it would be a good idea, just an interesting concept. Especially since it wouldn't be real people reaching out, but bots.

          Yea I don't really think it would be a good idea, just an interesting concept. Especially since it wouldn't be real people reaching out, but bots.

          1 vote
    2. [3]
      novac
      Link Parent
      I don't see why it would have to be directly through Facebook (Or any other social media platform). From what I could tell, the study only utilized photos, not captions or comments, so these...

      I don't see why it would have to be directly through Facebook (Or any other social media platform). From what I could tell, the study only utilized photos, not captions or comments, so these photos could simply be collected and fed into a program without ever having to use social media in the first place.

      1 vote
      1. NubWizard
        Link Parent
        True. And I will add that my view is that of ignorance and pessimism on the extraction of personal information through markers such as this. You get introduced to the internet and are cautioned on...

        True.

        And I will add that my view is that of ignorance and pessimism on the extraction of personal information through markers such as this. You get introduced to the internet and are cautioned on sharing too much personal information which you would think only applies to making posts/comments and possibly expression interest/liking posts on controversial subjects but now you have to worry what information is being extracted from the photos you share. With something like this, my mind jumps to when Facebook experimented on its users:

        https://www.makeuseof.com/tag/facebook-secret-experiments/

        What happened: Facebook data scientists manipulated the news feeds of almost 690,000 users, showing some of them more positive updates and others more negative ones. All to see how it affected the users’ moods.

        Or..

        What Facebook found out: During the three days of running this study, the researchers found that the friends of the users with positive updates were suppressing their negative posts and vice versa. If you post something positive on Facebook, one out of every 100 friends (who wouldn’t have otherwise, according to the study) will do the same within 3 days.

        Combining the information gleaned through analyzing photos with other information that is fed into Facebook could make an even more powerful, emotionally-manipulating advertisement platform. I don't know either if the fact they analyzed social media pictures has on the applicability of this research on utilizing the same methods for the uncurated pictures that are left unshared on your device.

        But like I said, I am all for this tool as something a mental health professional could use to help guide therapy and that's what makes this research really cool and interesting. I just hate the idea that something like it could be used in an unethical manner to generate some ROI.

        2 votes
      2. J-Senior
        Link Parent
        They analysed Instagram photos so that they could factor in like and comment counts. They didn't read the comments though.

        They analysed Instagram photos so that they could factor in like and comment counts. They didn't read the comments though.

        1 vote
    3. crius
      Link Parent
      I'm very pro technology advancement but I always follow the approach of human enhanced by technology. This tool could definitely help if it was something that crawl the web, can recognize who you...

      I'm very pro technology advancement but I always follow the approach of human enhanced by technology.

      This tool could definitely help if it was something that crawl the web, can recognize who you are (require you to actually be registered as an opt-in) and forward a notice to a medical professional of your choice.

      Parents/Guardians allowed to register their children as well. This would be probably the most common use-case.

      In this configuration, nothing is decided by an algorithm but still by a human trained person.

      About privacy, this is really not a technology problem but a "human" problem.
      If the company/government that implement this, decide to abuse it, there isn't much that technology (or developers) can do besides not use it.