Amy CahillComment

Digital Health and Behaviour 101 : What We Share

Amy CahillComment
Digital Health and Behaviour 101 : What We Share

It’s time we develop the ‘Sex-Ed’ for Social Media Use

I’m all for going to the pub for a drink after work and yes, I have an occasional tendency for that to lead to a 4am cab ride home. Moderation is accepted as key to enjoying this as part of our lives. Like sex ed, drug and alcohol education programs or learning a balanced diet is more than Vegemite on toast; social media is a part of our life that has a direct impact on our health and wellbeing.

Yes, we regulate alcohol and tobacco, but I’m not going to get into discussions on regulating social media because that’s a rabbit hole I haven’t had enough beer to explore. What I think we are missing is an education on its influence, it’s impact on our sense of self and health, and (at the risk of sounding like Healthy Harold) developing healthy habits. I just want to discuss how to empower individuals with enough information to make some of these decisions themselves.

Personally, a healthy amount of social media means no Facebook. I get questioned about it all the time. Probably because I’m a millennial in the tech industry and that doesn’t match the big data algorithms and stereotypes. I don’t want the world to delete their Facebook accounts, but I do want to start a conversation about our relationship with social media. I think the only way to start this conversation is to be very clear on the data we are sharing, how it is used, and openly recognising these platforms want to keep you on there longer. Watching the public outcry over a census form in Australia and now facial recognition, I think I’ve taken for granted how much we sign away when we click ‘Log in with Facebook’.

Here are a few of core concepts will make this easier to explain…

Psychometrics — The study of measuring your mental capacity and processes. Think IQ tests, Personality tests, Myers-Briggs etc.

The Big Five — The Big Five Personality Traits are one of the most widely used frameworks personality defining frameworks. Sometimes called OCEAN, the core traits are Openness, Conscientiousness, Extraversion, Agreeableness, Neuroticism. These traits have been empirically linked to things such as romantic relationships, academic performance, persuasion and learning styles. If you want to test your own check this website here.

Data — Everything is data. Everything you share and everything we can determine based on that information you share.

Big Data — simply, when a whole shit ton of data exists it can be mined for patterns, trends and prediction.

Facebook — Facebook have a huge network of partners buying targeted ads and leveraging the platform — offering content, quizzes, competitions and services in exchange for you sharing a certain amount of data to connect with their own. When I refer to Facebook in this post, I mean the platform and the network of partners accessing the data in and from it — everything from academic institutions, media companies, private consultancy groups, big corporates etc. etc.

You Are Not Unique

We’re all aware that we share a lot with Facebook, from our location, our acquaintances, our preferences on whether or not pineapple should be a pizza topping. We’re comfortable in sharing this data because it’s convenient, we get benefit from it and it’s a great way to share your life with friends and family separated by time, distance or necessity. The challenge is that we often don’t understand how much we’re sharing when we click like on a page, share a post or upload a new photo. Let’s look at just how accurately your Big 5 Personality can be predicted based purely on you Facebook Likes.

Based on about 90 page likes, FB knows your personality better than your friends.

150 and the parents who raised you from birth can’t compete

300 and they know better than your spouse

Like enough pages and the data from your profile understands you better than yourself, which for young kids still developing their sense of self is a particularly worrying stat. Particularly when insurers can begin targeting individuals exclusive to those personalities that are less impulsive or ‘sensation seeking’.

“But I like a whole bunch of weird and random stuff and follow politicians I don’t actually like”

Yes, but so does everyone else. Stats work with likelihoods not certainties and are pretty great at picking up outliers. If you follow enough people/places/things and then follow the patterns and trends across the literal billions of daily users, it turns out you can predict with a certain amount of confidence. Liking and following Donald Trump may not tell you whether an individual is pro-Trump or anti-Trump. Associating that with also following Fox News and The Red Pill tells a very different story to someone who also follows The New Yorker and Alyssa Milano. It’s why we can’t say that every follower of Lady Gaga is an extravert, but we can say that it is a strong indicator. When computers can process innumerable inputs, we really do become predictable. If you want to test your own, Cambridge University (not Cambridge Analytica) offer the service here. Bonus; they’re probably not going to sell you out.

But what about everything else you share? Your text posts and comments can be analysed for sentiment and linguistic style — based on writing samples alone we can predict gender, confidence and education. Write enough emails and we can begin to imitate your writing style for an auto response. Pictures or video for facial recognition, age, gender and emotion — even just the amount of photo postings can be correlated to your confidence and level of extraversion. Your location data provides a wealth of data around likely political leanings, socioeconomic status and key political concerns. Looking at anonymised location data, four random points are enough to predict who you are with greater 90% accuracy, something Uber leveraged for questionable use.

The effectiveness of data and analytics have been a topic of conversation well before social media, the classic case of a shopping rewards program picking that a teenager was pregnant before her family knew based on her shopping cart. The challenge is that today this isn’t just being used for advertising, or a mail out with some weirdly specific discounts. There are persistent rumours that our devices listen to our conversations to optimise marketing but honestly, I believe most platform providers don’t need to hear your conversations to predict that you may be looking at a fitbit, or need to organise insurance for a new car. Matching first- and third-party data to social data means that as soon as you look at a house in a new suburb on a property site, banks can proactively offer you a special personalised home loan rate based on the location, estimated size of loan, existing relationship with the bank, your credit score, are you in the age bracket to be buying a first home, or expanding for children…

Your Personalised High

It’s no different than driving education in schools around physical activity, drug use, diet, sex — focus on what the risks are, educate them on healthy habits, teach people how recognise when someone is struggling. Without going into the depths of the potential negative impacts of social media (short pieces available here, here and here) — I wanted to focus purely on how addictive Social Media is.

Social Media platforms are quite clear on the KPI’s they report to market. They want you on their platform for as long as possible, they count Daily Active Users, Average time spent on the site, where you access it from, how often you access the app — the more time you spend on the platform the more money they make and the more data you provide them. Personalised to meet your eerily specific preferences, the aim of the timeline is to win more of your time. If anyone can distinguish that from the term addictive let me know. In the tech industry, the value of the level of attention a push notification receives is well known

In lieu of trusting producers with moral decisions on consumption when it comes to vices, we take responsibility to educate citizens on the effects, risks and healthy dose. Before installing a time tracking app that told me I was spending 9 hours a week on Facebook, a couple of warning signs became obvious that I was wasting too much time on the platform.

If you wake up in the middle of the night, do you check your phone?

Is your phone the first thing you check when you wake up?

Do you have notifications turned on? For what?

When you have notifications on, do you unlock your device with some, most, all notifications?

How often do you check your phone without a notification?

Where this conversation sits squarely at the moment is teaching young people how to behave online but they’re not the only generation struggling. The person I know who has suffered with an unhealthy social media relationship, and caused a lot of pain because of it, isn’t a digital native — they’re a baby boomer. Social Media has exploded quickly — and like alcohol some of us are predisposed due to circumstance or personality to not take well to it. Or take too well to it and become reliant on it.

Paying for a Service

I once heard a Chief Digital Officer say that the worst thing Facebook could ever do is serve you better ads. Let me put this bluntly, if you have ‘digital’ in your title and are espousing the value of data, you cannot seriously defend that position. Facebook sell billions of dollars of advertising on the basis that it’s influential and has a unique ability to hypertarget individuals to ensure it reaches the right audience. The problem doesn’t begin and end with Facebook’s decisions, governance and management, but every partner that purchases the advertising space or that you share your information with in exchange for a promotion, quiz, better or just so you can login to their site and not need to remember another password.

Let me provide three cases (that don’t involve swaying the outcome of an election) that present moral questions on how this content can be leveraged.

  1. Targeting individuals in low socioeconomic areas, currently job hunting with accredited training courses? What about Multi-level Marketing companies? What about high interest pay day loans?
  2. Targeting teens in high socioeconomic areas with greater introversion and neuroticism, displaying greater anxiety with tutoring services? What about fitspo content? Or diet teas or pills?
  3. Identifying individuals that are likely LGBTIQ and targeting them with links for mental health support services? What about in regions where they are persecuted?

Who makes the decision about ethical advertising standards on how this can be leveraged? Can we trust the platform receiving the profits? Do we need to work toward public policy that protects our online identities? More importantly can anyone really manage that at the scale of over a billion active users per day across the globe? I don’t think any of those questions can be realistically answered right now — particularly in the short term. If we want to work towards ensuring the Social Media remains, or becomes, of primarily public good it’s important we as digital citizens start having this conversation.

You don’t need sentiment analysis to tell I’m obviously biased and cynical, but very willing and excited to have a conversation. Thoughts and perspectives are always appreciated.

Amy decided she wanted to be Batgirl at age five. It hasn't really panned out, and now she is a tech geek by day helping businesses and government agencies deliver better experiences to their stakeholders. After hours she's a sport fanatic, music lover and part time MBA student.

. This site is a just a collection of the stuff that she does as a result of being a bit weird. It's occasionally updated so her mother knows she's alive.