Thomas Malone is a professor of information technology and organizational studies at MIT. In his recent book Superminds: The Surprising Power of People and Computers Thinking Together, Malone presents a clear and innovative framework for understanding how human societies are organized, and a preview of how communication technology is changing those systems. Malone considers human organizations, from corporations to professional guilds, “superminds” that guide collective decision-making. He identifies five different kinds of superminds: democracies, hierarchies, markets, communities, and ecologies.
Malone and I had a sprawling discussion that revealed new ways of thinking about everything from the potential of blockchain technology to the tactics of the Me Too movement and the global drift away from democracy.
Your book is about how technology coordinates large groups. Have you given any thought to the new possibilities blockchain presents for that sort of coordination?
It’s not my main focus, but there are some kind of obvious things. Blockchain will probably increase the desirability of using markets for more transactions, instead of having them occur inside hierarchies of various sorts. It potentially decreases the cost of having transactions in the market, since you no longer have to deal with a manual process for resolving disputes.
A second point is that in the past, some organization needed to underwrite a currency, or someone needed to supervise a credit card system, because it needed to be something that everyone could trust. When banks issue credit cards, they assure the merchant who accepts the credit card payment that that they’ll be paid. But blockchain reduces the business opportunity for those organizations, or at least it changes the business opportunity, because blockchain greatly reduces or eliminates the need for some organization to play the role of trust mediator.
And the third point is, if we want to think more expansively about this, a very provocative example is the DAO. That’s an example of a new phenomenon where even inside what would have otherwise been an organization, much of the management can be done in an automated way, if you have the right kinds of trust-guaranteeing mechanisms. It appears to have not worked, but for reasons that aren’t deeply implied by the approach they took. It seems to me quite plausible that a somewhat similar approach with details changed could be successful.
Blockchain will probably increase the desirability of using markets for more transactions, instead of having them occur inside hierarchies of various sorts.
So we have these communication abilities, and they’re creating what you call superminds of people interacting and making decisions. What’s the role of artificial intelligence in mediating or facilitating these group interactions?
I think we probably overestimate the potential of artificial intelligence in general, and underestimate the potential of what I call hyperconnectivity: the ability to connect people in far larger groups and in rich new ways that were never possible before. Even today, that’s by far the most important way most people use computers. There’s not very much AI in most people’s use of computers.
Well, we’ve come to take it for granted so quickly. Ten or 20 years on, very rarely are you going to sit down and marvel at what an accomplishment Facebook or Twitter are.
I think you’re right. One of the reasons we overestimate the potential of artificial intelligence is because it’s easy for us to overestimate how quickly we’ll have human-level artificial general intelligence. Because it’s easy to imagine robots as smart as people—science fiction is full of them. But unfortunately, it’s much harder to create such machines than to imagine them.
With hyperconnectivity, I think it’s almost the opposite—it’s easier to create it than to imagine it. We’ve already created vastly large, hyperconnected groups of people. But we still don’t understand what we have created already, much less what’s possible in the future.
But A.I. can help people work together in new ways. If you’re working together in the marketplace, you can have an automated agent that bids and buys and sells things on your behalf. If you’re in a democracy you can potentially have an automated voting agent that knows enough about your preferences to know who you would want to vote for.
Another very promising thing A.I. can do is to help figure out which people are best for doing which tasks. Imagine a world where you would only do the work for which you are the best person in the world to do it, or the work which you most wanted to do.
In your book, you classify five different kinds of superminds, or social organization: hierarchies, democracies, communities, markets, and ecosystems. Does being more self-aware about how we organize ourselves offer advantages for how we plan for the future?
Absolutely. Most of us are part of many superminds, whether we think about that way or not—the company we’re part of, the markets we’re part of, the communities we’re part of. Thinking about it this way gives us a useful and interesting perspective on our life.
Let’s start with the most obvious example. Every company, every organization is a kind of supermind, right? Yet we don’t usually think of it that way. We think of it as a bunch of individuals. There’s a boss, and there are my friends I work with, and there’s this weird guy over in accounting, he’s pretty crazy. But most of us don’t think about the whole thing as an entity, as a potentially intelligent entity. And I think that alone is a powerful and useful way of thinking.
The second point is that I think the five types of decision-making superminds are a very useful framework for thinking about how to do things in the world.
Can you give an example of how the framework helps?
Suppose you believe, as many people do, that women are not treated as well as they should be in many workplaces today. And you want to change that. This list of five decision-making supermind types gives you a kind of checklist of possibilities to consider. If you want people to be treated better in your workplace, you could try to convince the people at the top of your hierarchy to make changes in the company’s policies that would cause that to happen. If you wanted to make a democratic legal change, then you should be doing political fundraising, or political demonstrations, or vote getting to try to change the laws about how women are treated in workplaces.
Still a third way of trying to make changes like that would be to do it using communities. And this is sort of what the Me Too movement has done. Communities make decisions through informal consensus based on shared norms and expectations. If you want to make changes that way, you don’t have to change any laws or any policies. You just need to change what people think is the right thing to do—you need to change the norms. And one way of doing that is by appealing to other norms, or people’s values, and working on the basis of reputations.
The thing that information technologies do very broadly is reduce the cost of making decisions in groups.
That example offers an interesting historical parallel. In a good sense, that movement has used public shaming as a tactic. That goes back to smaller, less technologically advanced communities.
Exactly. Shaming is something that in some sense only has meaning within the context of a community of people who share norms, and who can be shamed if they don’t follow the norms of their community. One way that norms can sometimes be changed is just by acting as if they’re different, and making people feel ashamed that they’re not with the program of the new norms.
At the same time, though, there are communities in the U.S. that don’t seem to be on the same page, and don’t seem to be experiencing shame around the same things.
Exactly. That’s one of the limitations of using communities as a way of making change. They only work when there’s a genuine community. And unfortunately, what appears to be happening in the United States today is that our community, which used to be more coherent across the country, is fragmenting on the basis of political beliefs, and maybe even cultural beliefs. If you want to do it for everybody, then you more or less need government hierarchy to do it.
But there’s another way of making the change, and that is to use markets. If you believe, as I do, that treating men and women more or less equally can make you more effective in the marketplace, then perhaps your best strategy would be to not tell anyone, and to do that in your own company. And then you will have a competitive advantage, if you can take better advantage of the capabilities of all your workers, regardless of their gender, than your competitors. That advantage will last until they figure it out. And eventually that would presumably spread through the market.
And my understanding is that there is research that at least large organizations are more effective when they’re more diverse.
I think it’s a complicated story. Some of our own research has found that small work groups are more collectively intelligent when they have more women. In that case, the main factor is what you could call social perceptiveness or social intelligence. There appears to be a strong relationship between that and the collective intelligence of a group, and at least the way we measured that in our study, it was already known that women score higher on that measure than men do. Though only by a little.
And I guess just for completeness, we could say the last decision-making method, ecosystems, is based on power. So if you have various ways of using your power, whether that’s physical power, economic power, or whatever, you can maybe make people do this, whether they want to or not.
So, to the point you made about the expanding role of markets. We’ve already seen that one of the downsides of this more individualistic labor market is that individuals can wind up with less power, less ability to protect their own interests. How can we make that transition better for more people?
That’s a question I wrote about in my 2004 book, The Future of Work. One of the possibilities that I think is very interesting, and that is surprisingly little-discussed, is guilds. If you have a bunch of people who are now independent workers, they don’t have a single employer to guarantee things like income security from month to month, or health insurance, or a place to develop your skills and be credentialed—or even a place to socialize and a sense of identity.
The idea is to set up independent organizations whose job it is to provide those kinds of services for independent workers. The term we ended up liking best was guilds, kind of harking back to the medieval craft guilds, but without the connotation of being a monopoly on a particular kind of work. There could be multiple guilds competing to supply the same kinds of services to the same kinds of workers.
For instance, you might pay a certain percentage of your income to the guild in good times in return for having a guaranteed minimum income in the bad times. That’s unemployment insurance, but rather than being provided by a distant, impersonal bureaucracy, like a government, it’s provided by your fellow guild members. It’s also a nice coherence of objectives, because if the guild is responsible for paying you when you’re not working, it’s definitely in their interest to help you find work.
I’m actually part of an organization for writers that fits some of those characteristics, Study Hall. Lately they’ve made some statements about employment practices that have had some impact. So it’s a little bit like a union, but not quite.
That’s very interesting. I do think unions are one place these guilds could come from. Some unions do some of these things already, and I’ve had conversations with unions that are interested in providing more of them. They can also come from temporary help agencies, or from college alumni associations. Ebay did some of these things for their power sellers.
Most of us don't think about the whole organization as an entity, and as a potentially intelligent entity.
Tech isn’t just fragmenting workplaces. We’re connected to people who are farther away from us, and with larger groups of people. There’s a concern that we’re just not designed, biologically, to be engaged with these huge abstract groups on a regular basis.
This is a big question, and a very interesting one, but one for which there’s not a simple right answer.
I think there is truth to the speculation that something about our physiological and cognitive makeup is probably well adapted to the world in which we evolved genetically, the world of our hunting and gathering ancestors. In that world, most people lived in bands of roughly 20 to 50 people.
Robin Dunbar, the anthropologist, has argued that about 150 is the number of people that we can keep track of and maintain some sense of relationship with. And I think we probably evolved to fit well in such groups. One implication is that in organizing large groups, you should have people in subgroups no bigger than that, and probably even smaller.
I’ve heard that number in discussions about social media.
Certainly. I think another dimension in play here is the richness of the interaction. With today’s technologies, there is still a kind of sensory richness you get when you are face to face with someone that you don’t get when you’re just talking to them electronically.
Now, notice I said that’s with today’s technologies. As the technologies get better and better, the difference becomes less important. You and I are talking by phone. We can’t even see each other’s faces, but we hear each other’s voices with very good fidelity. Neither you nor I thought that this conversation was important enough to fly from New York to Boston, or vice versa. If the only alternatives were having a face-to-face conversation or nothing, we probably wouldn’t have had the conversation.
So that means there are a quite large number of communication interactions that are enabled well enough by this new technology that we’re happy to have them, and that wouldn’t be possible without it. But I do think there is a risk of lacking something with the somewhat impoverished nature of communication—people have different desires and tolerances for social interaction. For myself, I’ve found that telephone interactions are better than nothing, but not as good as face to face interaction, for preventing the feeling of isolation.
I think we probably overestimate the potential of artificial intelligence in general, and underestimate the potential of what I call hyperconnectivity.
When there are conflicting conclusions between these different decision-making systems—between a hierarchy and a community, for example—is there any consistency to how those conflicts get resolved?
Almost all the groups we now have actually include more than one of those decision-making supermind types. Most companies, for instance, have a hierarchy that’s pretty obvious, but they also have a community, a corporate culture. It’s a set of norms and reputations that are a very important part of how the company functions. And many companies also have some kind of market within them, and some make decisions with a kind of democratic process, whether they have formal voting or not. I think there’s often aspects of all five of these different types of decision-making superminds in play in a given group or a given situation.
But in some sense, ecosystems are always in the background. If there is a community that has certain norms and functions effectively making decisions based on these norms, there’s always an ecosystem waiting in the background. If somebody who’s very physically strong or has nuclear weapons appears, they can overpower any other kind of decision-making type, by pure power.
You could argue that that has happened in the United States in the last century or so. A lot of things that used to be done in communities are now done by markets, because if the market is more appealing than the community, then it has more power. People chose to buy and sell things instead of doing them as communities, so the markets beat the communities in the deeper ecosystem.
You set me up for my final question. Right now, you see people advocating for different modes of organization, different superminds. There’s been a lot more interest lately in political libertarianism, which in some ways is sort of the ecosystem model. And especially in other parts of the world, there’s a drift toward hierarchies in government and away from democracy. Is this all part of a larger process?
I think it’s absolutely all part of the larger process. If you look over the last few centuries or millennia, markets have become much more important, and democracies have also become much more important. One of the things that you can predict is that the decision-making types for which technology reduces the importance of their weaknesses should become more common.
And the thing that information technologies do very broadly is reduce the cost of making decisions in groups, which requires communication, and things like counting votes. So if a technology reduces the costs of group decision making, it means that democracy, and at least certain kinds of markets, are likely to become more common in the future.
That’s an optimistic note to end on. Thanks for your time, Dr. Malone.
Thank you. Take care.