How They Do It & How It Hurts

Users think that they have control over their social media habits and the information being fed to them, including news and suggested webpages, are coming to them organically. But, unbeknownst to them, the framework is calculated. The US Deep State, for instance, helped to develop social networks. Sergey Brin and Larry Page developed their web crawling software, which they later turned into Google, with money from the US Defense Research Projects Agency. Referring to the Massive Digital Data Systems, the CIA-funded Dr Bhavani Thuraisingham confirmed that “[t]he intelligence community’s MDDS program essentially provided Brin seed-funding.”

Consider how the technologies were commercialised. “Growth” means advertising money accrued from sites visited, content browsed, links clicked, pages shared, etc. “Growth hackers” are described by former Google design ethicist Tristan Harris as “engineers whose job is to hack people’s psychology so they can get more growth.” Designers build applications into software that manipulate users’ unconscious behavioural cues to lead them in certain directions.

To give an example: The feel-good chemical oxytocin is released during positive social interactions. It is likely stimulated when social media companies send an email alert that family have shared a new photo. Other human foibles include novelty-seeking (for potential rewards) and temptation (fear of missing out or FOMO). These are linked to the feel-good chemical dopamine. Rather than including the new family photo in the email, the email is designed with a URL feature to tempt the user to click the link which directs them to the social media site in order to see the new photo. The chemical-reward response chain is as follows: family (oxytocin) → novelty/new photo (dopamine), temptation to click/FOMO → reward from positive social interaction after clicking and seeing the new photo (oxytocin-dopamine stimulation).

This convoluted chain of events is designed to sell the user’s attention to advertisers. The more time spent doing these things, the more adverts can be directed at the user and the more money for the social media company. Harris says “you are being programmed at a deeper level.”

In addition, tailored psychological profiles of users are secretly built, bought from, and sold to data brokers, like Experian. User behavioural patterns feed deep learning programmes which aim to predict the user’s next online move according to their personal tastes and previous browsing patterns. The more accurate the prediction, the more likely their attention is drawn to an advert and the more money social media firms accrue. Says former Mozilla’s Raskin: “They’re competing for your attention.” He asks: “How much of your life can we get you to give to us?”

Instagram was developed in 2010 by Facebook as a photo and video sharing service. It is used by a billion people globally and, unlike the teen-loving Snapchat, is used mainly by 18-44-year-olds. Instagram falls into the so-called “painkiller app” category. One designer explains that such apps “typically generate a stimulus, which usually revolves around negative emotions such as loneliness or boredom.”

Snapchat is a messaging app designed in 2011 that stores pictures (“Snaps”) for a short period of time. The app is used by 240 million people per day. Unlike YouTube, most of whose users are male, the majority of Snapchat users are female. Only 17 per cent of users are over 35. Its model is Snapstreak: a tracker that counts the days since the user replied to the Snap. Designers built FOMO (noted above) into Snapchat. The longer the user’s non-reply, the greater their credit score decline. This can lead to addiction because, unlike Facebook, Snapchat tags are “strong ties” (e.g., close friends, family), so the pressure to reply is greater.

In addition to the harmful content of social media – sexualised children, impossible and ever-changing beauty standards, cyberbullying, gaming addiction, loss of sleep, etc. – the very design of social media hurts young users. We all need to love ourselves and to feel loved by a small circle of others: friends, family, and partners. Young people are particularly susceptible to self-loathing and questioning whether someone loves them.

The introduction of social media has been devastating. A third of all teens who spend at least two hours a day on social media, i.e., the majority, have at least one suicide risk factor. The percentage increases to nearly half for those who spend five hours or more. A study of 14-year-olds found that those with fewer social media likes than their peers experienced depressive symptoms. Teens who are already victimised at school or within their peer-group were the worst-affected.

Divided & Conquered

Another feature built into social media is the polarisation of users along political lines; a phenomenon that mainly concerns people of voting age. One of the many human foibles exploited by social media designers is homophily: our love of things and people similar and familiar to us. Homophily makes us feel safe, understood, validated, and positively reinforced. It stimulates feel-good chemicals and, in social media contexts, is exploited to keep us inside an echo-chamber so that our biases are constantly reinforced, and we stay online for longer. But is this healthy?

Referring to Usenet group discussions, the lawyer Mike Godwin formulated the Rule of Hitler Analogies (or Godwin’s Law), which correctly posits that the longer an online discussion, the higher the probability that a user will compare others to Hitler. The formula was a reflection of users’ lack of tolerance toward the views of others.

A projection published in 2008 asked if people will be more tolerant due to the internet. Nearly six in 10 participants disagreed, compared to just three in 10 who agreed. In many ways, the industry specialists were fatalistic. Internet architect, Fred Baker of Cisco Systems, said: “Human nature will not have changed. There will be wider understanding of viewpoints, but tolerance of fundamental disagreement will not have improved.” Philip Lu of Wells Fargo Bank Internet Services said: “Just as social networking has allowed people to become more interconnected, this will also allow those with extreme views… to connect to their ‘kindred’ spirits.” Dan Larson of the PKD Foundation said: “The more open and free people are to pass on their inner feelings about things/people, especially under the anonymity of the Internet – will only foster more and more vitriol and bigotry.”

Users can artificially inflate their importance and the strength of their arguments by creating multiple accounts with different names (“sock puppets”). Some websites sell “followers” to boost users’ profiles. It is estimated that half of the Twitter followers of celebrities and politicians are bots. Gibberish-spewing algorithms have been programmed to write fake reviews on Amazon to hurt competitors’ sales. In at least one case, a pro-Israeli troll was unmasked posing as an anti-Semite in order to give the impression that anti-Semitism is rampant online and thus users should have more sympathy with Israel. Content creators increasingly find themselves de-platformed because of their political views while others’ social media accounts are suppressed by design (“shadow-banning”).

In the age of COVID, disinformation on both sides is spread: the severity of the disease, efficacy of vaccines, necessity of lockdowns, etc. As with US politics, Brexit, climate change, etc. neither side wants to talk rationally and open-mindedly with the other. The very designs of social media make this very difficult.

It should be emphasised that some social media are designed to create echo-chambers, and others are not. Cinelli et al. studied conversations about emotive subjects – abortion and vaccines – and found that while Facebook and Twitter show clear evidence of the echo-chamber effect, Reddit and Gab do not. Sasahara et al. demonstrate that due to users’ need for validation, when likes and friendships are withdrawn the network tends to descend into an echo-chamber.

Conclusion: What Can We Do?