Ted Littleford
A warning, and call to action, as tech giants prepare to testify before Congress.
The protest at Kent State is being live streamed and millions watch as the National Guard troops take their positions. Suddenly, shots ring out, and four students fall. Within seconds, several hundred thousand tweets and multiple storylines cover the gamut of political attitudes.
A rumor starts that a group of anarchists shot the students from positions near the National Guard. Some of the dead students are identified in minutes, and within an hour some right wing groups have connected them to left wing terrorist groups and suggest that they were shot because they were about to set off a bomb. Some left-wing groups publish locations of other groups of guardsmen and urge their followers to confront them.
A group of 12 Guardsmen are surrounded by 1,500 protesters near the administration building …
Now imagine a scenario like this one, which could take place anywhere in the world:
Marnie Jones, a high school junior, has been dating Robert Smith for about a year. She discovers that she is pregnant.
On the same day, she sees postings of Robert with another girl, and it’s clear that they are involved. She texts him and he texts back that he is breaking up with her. She tells him that she is pregnant, and he texts back that the baby can’t be his, accusing her of being unfaithful.
Later in the afternoon she sees that he has posted hostile messages about her and naked pictures of her. Later in the week, she is photographed visiting a planned parenthood office and the photos are also posted with the tag #slut. She gets a message on Facebook from an anti-abortion person in Australia accusing her of being a baby killer. …
Such specific scenarios might not come up when Congress undertakes an historic grilling Wednesday of the leaders of the country’s tech giants.
But they underscore how much we as a society, and we as individuals, have changed as a result of the predominance of social media in our lives, unleashed and refined by some of those same companies.
The hearing will explore, anong other questions, whether, since social media is so powerful, its control should remain in the hands of a single person, like Mark Zuckerberg, or a small group of companies.
A more fundamental question is how social media has changed us as human beings and changed how we act toward each other — not for the better. And what we can do about that.
That quest begins with an understanding of how we arrived at this place.
Our brains did not evolve to function well with online networks, and we are experiencing the consequences of their indiscriminate growth.
Social media has been around long enough that we can be sure that it is having serious negative effects on emotional health, politics and the culture. It’s an addiction, and like many other addictions, one with bad consequences.
Before nuclear weapons, we had developed various ways to manage the levels of violence we could perpetrate. But then we invented the ability to commit violence beyond the scale of our reason — a click of a button could annihilate a city, a civilization. Therefore, we had to learn how to think about nuclear war, and manage conflict in a world in which our technology was capable of ending us. Our thinking had to evolve.
Similarly, we’ve been unprepared for the scale and velocity of social networks. There is simply no precedent for a continuously available social network capable of distributing arbitrarily large volumes of content to over 2.6 billion people, all of whom are in principle connected to each other.
Our brains did evolve at much lower scales of social and informational complexity, bands of a few hundred up to loose groupings somewhat but not significantly larger. Imagine the typical number of messages and social interactions. Imagine the content of those interactions. Compare in your mind to what occurs on Facebook. These two environments are in every sense qualitatively different. We have taken a mind that evolved under one set of circumstances, transported in effectively a blink of an eye to this alien world.
There is not a doubt that we have overloaded our capabilities — cognitive and emotional — as we’ve shifted our social references online. This is one of the great issues of our time; this transition has serious consequences, evidently for our elections but, in fact, for every aspect of our lives. We might imagine ourselves (especially in the United States) as perhaps influenced by our social networks, but fundamentally independent. This conceit is false. Our cognitive and emotional lives are profoundly mediated by our social networks — the way we feel, how we think, what we think, what moves us to action are all moved by the social currents. We developed in rivers but now find ourselves in an ocean.
There were social networks long before computers. These social networks were key to our evolution and shaped us. We learned how to manage important functions — how we could learn to trust, how to modulate influence, how to form and manage groups. We learned, however imperfectly, to manage a certain level of complexity. The modern platform of social networks has seriously undermined those adaptations.
Modern social networks will not go away. We are looking for ways to mitigate the bad effects while we take advantage of stunning new possibilities. The government has talked about regulation, and companies like Facebook tout technological solutions like AI. Let’s start towards a solution by trying to understand more deeply what has gone wrong.
Consider the three pillars of social network function: trust, influence and group dynamics. All three have been radically affected by the new medium.
Trust
Trust is the prerequisite. In simpler times, it was likely dominated by identity (are you from the same group?) and through relational extension (trust if the person is trusted by people you trust).
There are many dynamics underlying trust but just sticking with these two primitives, it is easy to see how modern networks make everything so much more complicated. You ‘meet’ so many more people.
There are so many different kinds of groups — Google now indexes 620 million groups on Facebook. With connectivity on the social network (six degrees of separation), we often have the strange notion of sort of knowing a lot more people. A simple trust decision goes from being a relatively simple recognition problem to something much more complicated and harder to pin down.
We also have many more of these decisions because there are so many more interactions in virtual space. The cost of interaction is so much lower. I can say a quick hi to someone in Zanzibar as easily as I can to my daughter in California.
Influence
Once there is some level of trust, then influence becomes important. Again, there is the presumptive moral superiority of the independent thinker, but our greatest accomplishments come from collaboration and this requires the ability to be influenced. Had our ancestors not been “influenceable,” we would have had a lot of trouble. Individuals that stood adamantly alone had many fewer offspring. Influenceability evolved as a necessary trait.
Researchers have identified many influence mechanisms. Cialdini’s famous list:
Reciprocity: I do something for you and then you owe me.
Liking: You like me so I influence you
Consistency: You try to align your statements/beliefs/actions.
Social Proof: we do tend to jump off bridges if our friends do.
Authority: Yes, it matters, a lot.
Scarcity: People feel compelled to act when they perceive scarcity.
We could go through each of these and show how differently they operate (and powerfully) online, and how that yields opportunities for bad actors. But let’s focus on two facts:
• Volume (and hence cognitive load).
It’s impossible to know how many influence messages a typical person received in the thousands of years before online networks. Let’s say it was a few dozen. (I’m talking about messages explicitly targeted to the user to elicit some thought or behavior.)
Leaving aside ads which may be in the thousands across all forms of media (and take their cognitive toll), a typical user is going to see orders of magnitude more than a few dozen explicitly and implicitly influence-oriented messages. This certainly includes political stories sent to prove a point, arouse a passion or enforce a perspective.
The sheer weight of all of this influence messaging has an effect on our limited ability to take on cognitive load.
• Social proof.
In the simpler pre-network world, how many examples of social behavior did we see? Television upped the number, but still it was trivial compared to the number available in five minutes of social media. From a controlled pipe, distributing some limited content, we now have an infinite well of visible behaviors.
For fun, I searched on YouTube for “dropping watermelons” and didn’t bother to count the many hits. I looked briefly at the related videos and saw that I could jump to “firecrackers in a toilet.”
Every attitude, every behavior has some expression on social media, and it drifts around drawing people in sometimes for the banal and sometimes for the truly dangerous. Again, there is no precedent.
Group Dynamics
wearesocial.com
This is a huge topic, so let’s again focus on an illustrative point. In the physical world, groups have natural limits to membership and limits on how members interact and are observed. In the online social network world, there are no such limits. If you publish a comment, your entire group will see it instantly and react in ways that suit their interests in being a certain way.
The online groups are like panopticons that drive groups towards purity and extremism. In some sense, the group has a hyperactive immune system that reacts to incipient deviance as though it were a fully developed rebellion.
Real world cults try to replicate this effect, but in the social network world it is fundamental to the platform.
Our brains are doing their best to operate in this new world for which they were not designed. Researchers are trying to measure consequences. But it’s clear that important fundamentals of how social networks have worked historically are being undermined. Scale and hence cognitive load are critical to this change. We are overwhelmed individually and collectively. We have not yet figured out how to manage to live in this world we have created.
The above discussion is intended to be illustrative. We can and should look more deeply at how the fundamentals have shifted and how we can adapt to this. It’s not clear how a government can effectively and ethically regulate social networks. “AI” is not an answer. It’s a technological hope. Let’s first try to understand the problems we are trying to fix. The stakes are extremely high. Many people feel that something is fundamentally wrong with the country, the polarization seems uncontrollably vicious. Social networks and the pathologies that they currently enable are very much part of the reason.
Jeffrey Davitz is a scientist/entrepreneur who lives in Branford. He has worked for decades in AI and social computing.