Can We Amplify the Good and Contain the Bad of Social Media?

In his new book, The Hype Machine (Currency, 2020), MIT Sloan professor Sinan Aral takes on the greatest communications force of our lifetime: social media. Aral, who directs the MIT Initiative on the Digital Economy, provides an insightful and level-headed analysis of the power, peril, and potential of Twitter, Facebook, Instagram, and other social platforms for both organizations and society. The work is the result of some two decades of research he has dedicated to social media. He spoke with MIT Sloan Management Review’s editor in chief, Paul Michelman, about the dual-edge sword of what he terms “the hype machine.” What follows is an edited version of their conversation.

MIT Sloan Management Review: Is there an approach to social media marketing in which we can exploit the power of the media without exploiting the people who are using it?

Sinan Aral: Yes, I think absolutely. The interesting thing about marketing is that it is a technology like any other. When we think about marketing to promote voting or to stop dirty-needle sharing or to get people to socially distance or wear masks during the coronavirus pandemic, we don’t think of it as manipulative or in some way malign. We think of it as persuasion to promote positive behaviors.

Marketing is a tool that can be used for good and evil. And that is part of the message of the book. This tool, the hype machine, the social media industrial complex, is powerful. It’s used for election interference, and it’s used to market the use of condoms or to prevent the spread of disease. And so the point of the book is that we can achieve the promise and avoid the peril, and the book tries to explain how we can do that.

What are some of the most important principles that differentiate using social media for good, or at least neutral, purposes versus using it for nefarious purposes?

Aral: The devil is in the details of what you are using it for. Businesses can use social media to make consumers aware of their products and services, to be targeted in their communications so that they are speaking to the people who would get the most value out of their products and services. I frequently see, for instance, advertisements for solutions to diseases or to conditions that people have. And I think to myself, “Wow, if the people who actually had these diseases and conditions could hear these messages, how much better their life might be.” Targeted messaging can really help get the right message to the right people, and social media can be incredibly well targeted. That’s part of its promise, when it’s used correctly.

However, election manipulation is equally targeted — for instance, to suppress the votes of Black voters in the United States. That is targeting for nefarious purposes, and it’s something we saw explicitly in the 2016 U.S. presidential election. The hype machine itself is the tool, and the way it’s used can be for good or bad.

It’s very tempting, or it could be, for someone who doesn’t have ill intent to still look at some of the practices being used by more nefarious players and want to copy them.

Aral: What gets directly to your point are the two questions that I’m asked most often in my role as an educator and a scientist at MIT. One, did Russia change the results of the 2016 U.S. presidential election through its election manipulation on social media, and will it change the results of the 2020 election? And the second question is, how can I measure the return on investment to my social media investments?

And the interesting thing that I realized in writing the book is that the answer to those two questions is the same. There is a unity about social media that, whether it’s used by the Russians to affect elections or by businesses to promote their products and services, the same process can be used to optimize its persuasive power to generate engagement, insight, and action, and those processes are described in the book. Then another part of the book describes how we can limit the nefarious uses while promoting the positive uses, and that gets into the use of what I call the four levers: money, code, norms, and laws.

It seems like laws are not, or have not been to this point, a substantial part of the conversation. Do we need more regulation?

Aral: Yes. I think that regulation is a big part of the answer, and there are a number of different areas of regulation that need to be examined in detail and with a scientific approach. The ones that immediately come to mind are antitrust, which is front and center today. Does Facebook, for instance, need to be broken up? Then there’s privacy. Do we need federal privacy legislation in the U.S.? We have the GDPR [General Data Protection Regulation] in Europe. And we have China’s approach to privacy, which is the exact opposite, where the state can basically read and see all messages and data.

The regulation of free speech and hate speech, or the boundary between free speech and harmful speech, is another one. Political speech and political advertising on social media: Should it be labeled? Should it be regulated like political speech on television?

And fake news. The regulation of fake news is worrying, because in authoritarian regimes, such laws are used to suppress minority opinions. We see the case of Maria Ressa in the Philippines with the Duterte government using regulation to silence journalists. And there are countless other examples around the world. Each of these questions is its own Pandora’s box, but each of them, in my mind, has a path to a better future within it.

What about the role of the corporate sector? I don’t know if the Facebook ad blackout did any good, but it’s an example of some loosely organized private-sector institutions presumably at least trying to exert a positive influence.

Aral: Absolutely. As I mentioned, there are four levers: money, code, norms, and laws. The Stop Hate for Profit movement, and the Delete Facebook movement before it, really gets at money and norms together. We as businesses and citizens need to express the values we would like to see espoused within the social media industrial complex. We can espouse those values through movements like Stop Hate for Profit and affect the business models that the platforms use to create profit.

It demonstrates one tool that we have in our arsenal for steering social media toward promise and away from peril, which is to pressure businesses to change in the ways that we would like to see them change.

How transparent and public should companies be about the way they use social media as a marketing tool, what information they gather, and what they do with it?

Aral: Transparency is a very important piece of the puzzle, and I believe that it must begin with the platforms themselves. We need a lot more information as users about the provenance of the information we’re seeing, about the veracity of the information we’re seeing, about the context of the information we’re seeing.

Take fake news. When we consume food we buy from the grocery store, it’s extensively labeled. We know how many calories it has; whether it has trans fats; how many sugars it has; whether it’s produced in a facility that also processes wheat and peanuts, if you have allergies. All of that is labeled. But when we consume information on Facebook or on Twitter, it doesn’t have anywhere near the amount of labeling in terms of where the information comes from and what’s contained in the information.

Also, the platforms need to provide data to scientists and researchers so that we can understand the effects that social media has on our democracy, on our public health, on our elections, and on our economy. That kind of data transparency — and I outline many different ways we can achieve it in the book — is essential to understanding how the hype machine works under the hood and, therefore, how we can steer it toward a better future.

Now, in terms of businesses, I think that more transparency is also better for understanding how businesses use social media, but I think that the transparency really has to begin with the social media platforms themselves. That’s where it’s most needed and most lacking.

Let’s talk about the political echo chamber. How different is the power and influence of social media from the sway that newspapers once held over public sentiment? Yellow journalism was an incredibly powerful force. One of the things I did in preparation for this interview was to rewatch Citizen Kane for about the 50th time. And there’s that great line from Charles Foster Kane: “People will think what I tell them to think.” Now, is what Cambridge Analytica contributed to the 2016 elections any worse or more powerful than William Randolph Hearst and Joseph Pulitzer leading America into a war with Spain at the end of the 19th century?

Aral: I do think it’s potentially worse and certainly very different, for a variety of reasons. One is that things are much more hypertargeted. So with newspapers, with television, it was transparent what the messages were, and everyone saw all the messages. But now you can bifurcate the population into many, many, many small groups and send them messages that other people may not even be aware that those small groups are receiving. You can see attempts at manipulation or persuasion at the micro level but aggregated and coordinated to the population level, where nobody sees all of the other persuasion attempts that are going on. That’s very different than newspapers or television, and potentially very dangerous.

The other thing that’s very different is the algorithms. News feed algorithms take in what we like and engage with and give us recommendations for what to read next. Friend-suggestion or “people you may know” algorithms take in who we’re connected to and what we like in order to suggest who else we should follow. Those algorithms are going to perpetuate certain path dependencies in society and in the hype machine itself.

This is completely different than newspapers or television, and we have to treat it differently. We have to think about the solutions and our use of it very differently. In the book, I go into in detail on what I call the hype loop — the feedback loop between machine intelligence and human decision-making that generates the outcomes we see from the social media industrial complex. Understanding that hype loop is critical to understanding what we do next.

To participate in the hype loop, and to really take advantage of the notion of micro-influence, you have to have both the data and the intelligent technology to make use of it. What does that suggest about the types of organizations that will realistically be in a position to engage with the power of micro-influence?

Aral: The more data-savvy and rigorous organizations are the ones that will be more effective at using the hype machine to achieve their missions. But the process of doing that is actually quite simple, and I spell it out in very simple examples in the book — how to rigorously and scientifically measure ROI, and how to create an integrated and optimized social media marketing program. And those examples are brought down to the level of how small- and medium-sized enterprises, without big marketing budgets, can use it, and how very large organizations — for example, I talk at length about Procter & Gamble — can also use it.

Is the data and analytics technology sufficiently commoditized so it’s not the exclusive domain of big players?

Aral: Yes. The great thing about social media is it is an essentially democratizing voice, so anyone who is committed to learning how to use it can use it to get their message out into the world and can be persuasive with it. But it does take a commitment to a scientific approach to do it well.

I want to round out this conversation by talking about the nature of influence and influencers. Do I have to be open to being influenced in order to be affected by influencers? Or all we are affected equally if we’re exposed equally?

Aral: This is a really interesting question. At the MIT Initiative on the Digital Economy, which I direct, and in the Social Analytics Lab within the IDE, we have done a lot of studies about social media influence over the years. We’ve published them in Science and Proceedings of the National Academy of Sciences and so on. The research shows that there are differences between people who are influential and those who are susceptible to influence: People who are influential tend not to be susceptible to influence. People who are influential and have influential friends tend to be more influential than people who are themselves influencers but who surround themselves with noninfluential people.

Do influencers have a shelf life?

Aral: We’ve seen some influencers come and go, and we’ve seen others with tremendous staying power who remain influential year after year. If you had to pinpoint one thing that separates the two, it’s the ability of the perennial influencer to remain relevant and adapt to changing times.

I think that is a lesson for all of us. Businesses need to remain relevant. Opinion and thought leaders need to remain relevant. Educational institutions need to remain relevant. We went from a pre-COVID world to a post-COVID world in a matter of five months, and now everything is different. The people who are adapting best are the people who are learning by doing and doing by learning.