The inventor of the Facebook Like: “There’s always going to be unintended consequences”

The Facebook Like is a load-bearing wall in our digital lives. Since 2009, it has paved the way for a culture of instantaneous acknowledgement, sans words, sans pictures, sans all but a click. It has wound outside of Facebook’s bounds, into the hearts and notifications of every social media app. But it all started with a simple thumbs up.

In London, in a hotel bar, the Like button’s creator tells me it was intended as a nudge; a funnel; a path of least resistance. “We want to create a world where people uplift each other and promote positivity,” says Justin Rosenstein, former engineer at Google and Facebook. “So how can we make positivity be the path of least resistance?

“Comments mean you can say whatever you want, but having the Like button be this primary call to action means you bias people more towards structuring their communication around positivity.”

Like directing a downhill stream, Rosenstein and his fellow engineers aimed to forge a system that guides people into certain behaviours. Beyond Facebook’s grammar of interaction, Rosenstein has had a hand in developing myriad apps, from Google Docs to his latest venture, workflow toolset Asana. In each case, decisions had to be made around which buttons and menus would lead the user through which actions.

“Humans are very malleable,” he says. “It’s like architecture. Architects will understand that [one] structure will cause people to flow while certain architectures will cause people to feel isolated. Others will cause people to come together. When you’re architecting software, you’re creating similar dynamics where you can result in people exhibiting different kinds of behaviours in their lives unconsciously.”

But what happens if those systems bring about behaviours that you didn’t plan for? I ask Rosenstein what he regrets about the Facebook Like button. “When I walk around and see people staring at their phones often it’s because they’ve taken out their phones to look at Facebook notifications,” he says. “That’s something I feel is not going in the right direction for society.

“Whenever you create things there’s always going to be unintended consequences. I don’t think that’s a reason to not create things; to get paralysed. But I do think you have to be very careful and sensitive about what those consequences are, and keep watching and making appropriate changes so that you’re continuing to guide things in the right direction.”

(Above: Justin Rosenstein)

Do these consequences extend to the political? The 2009 rollout of the ‘Like’ widely predates 2016’s debates around echo chambers and purposeful misinformation. It has been joined with a wider palette of reaction emojis, but is the ‘path of least resistance’ towards positivity fit for purpose at a time when Facebook is, arguably, more of a publisher than a social network?

“As Facebook becomes more of a political platform, does it need more tooling to be appropriate for that? Yeah, definitely. It’s a question of how you avoid filter bubbles. One thing that’s happened with the Like button is that it’s good to know what people like, but it’s a mistake to always show them what they like.

“Sometimes we should be delivering content we don’t think you yet like, because we want to expose you to new information,” Rosenstein adds. “The ranking of a newsfeed is a very important problem that warrants a lot of thought.”

Rosenstein continues that software companies have a “huge responsibility” to think about the kind of behaviours they are generating in the world. He pitches this as something his current venture, Asana, is conscious of. Asana’s productivity toolset may be a far cry from a public-facing social network, but here too there are architectures that ripple down to behaviours. According to Rosenstein, Asana has been built with a fundamental bias towards accountability.

“It biases you towards thinking in terms that lead your organisation towards accountability, transparency, clarity. Getting people to slightly change the grammar of how they interact with each other, it’ll have all these emergent effects that are really powerful.”

It’s hard to tell how successful this alleged ‘path of least resistance’ towards accountability is, but it’s an interesting concept when you consider the software could – at least in part – influence the work of governments. I’m told that the US city of Providence, Rhode Island “runs on Asana”. If it were to expand across governments and security services, could the architecture of a humble work management tool influence the behaviours of major institutions, even nations?

(Above: City Hall for Providence, Rhode Island)

It sounds far-fetched, but this level of impact is very much on the Facebook Like creator’s agenda: “All of humanity’s big challenges are solvable with sufficient collaboration,” says Rosenstein. “We have enough food to feed everyone, we’re just not distributing it well. We have the ability to develop technologies for sustainable energy, there are just a lot of problems in the way. I’m not saying Asana can single-handedly get the world to collaborate, but I hope in ten years the world is much more collaborative and Asana is a core piece of technological infrastructure.”

He goes further. Rosenstein tells me he envisions an “omniscient AI”, embedded in something such as Asana, intelligently managing people’s workloads; giving them things to do, ensuring resources are used in the most efficient way. From there it’s only a hop, skip and a jump into a pan-global utopia.

“I would love to see that at a planetary level, where we see, as a species, we have this one mission of creating a world that works for everyone,” he says. “Where you can zoom in and see all the different problems that need to be solved in order to create that world. All the different organisations working to solve that problem.”

If the Facebook Like was built as a way to bias people towards positivity, Rosenstein’s plans for the future encompass vast, pervasive, AI systems that bias the world’s populace towards collaboration. Our tendency towards war is in the way of that, he admits, but the survival of our species ultimately depends on “a psychological transition from me to we”.

Grand thinking indeed. Whether or not our future lives are directed by a benign, omniscient AI, it is ultimately the systems – the human-authored systems – that steer our thinking. Will Rosenstein’s utopian outlook prevail? Will these structures guide us to a blazing world of positivity and accountability? Or will those troublesome “unintended consequences” always get in the way?

Rosenstein says that, if he were still working for Facebook, he would explore buttons that not only signaled things users liked, but that they considered the best way to spend time. He’d love to see things orientate more towards value.

“Liking may be too simple an emotion.”