Aside from your friends, colleagues and that meme-loving relative, the majority of your fellow social media users are complete strangers. And no matter how much time you spend online, it’s a remarkably tough task to gauge another person’s character -- especially when a cloak of anonymity can make it impossible to find sources of harassment or questionable information.
Biz Stone has a firsthand understanding of how the social media landscape has evolved over the last decade. In 2006, Stone co-founded Twitter with Ev Williams, who went on to found Medium, as well as Noah Glass and Jack Dorsey, who is the company’s current CEO.
Three years ago, Stone co-founded Jelly Industries, of which he is CEO, and he launched its product, Jelly. Billed as “a search engine for busy people,” the Q&A app allows users to ask questions anonymously about any topic. Stone and his co-founder, Ben Finkel, aimed to develop a community that would emphasize nuance and foster empathy.
To that end, this week, Stone announced a new feature for Jelly: a button that shows others whether you are a trustworthy source of information.
“Trust is earned on Jelly. It might be earned on our system because a person is particularly helpful to you, or maybe you are already friends (or virtual friends) with someone, so you trust each other automatically,” he explained in a post on Medium. “Trust is already starting to happen across the Jelly service. Helpful strangers are connecting via Trust. A network of trust is a beautiful thing.”
Entrepreneur spoke with Stone about the inspiration behind the new feature and how he is building his corner of the internet.
This interview has been edited for clarity and brevity.
Why was the Trust feature something you wanted to build?
We have a significant number of people who have signed up to answer questions. They all share at least one thing in common -- they enjoy helping other people on Jelly. We created the Trust model as a simple way for people across the service to acknowledge one another, to connect to one another and to discover one another. So begins the network of trust.
Also, when people get answers from Jelly, they should be able to assess quality. If you get an answer to a question about your kids, and the answer comes from someone with a high Trust count and “parenting” is one of the Topics they are most helpful in, then you can get a pretty good sense of the answer quality.
Why was it important to implement it now?
Trust is a big issue on the web these days. Implementing a system-wide feature enabling people to earn trust from other people based on their contributions, generosity and helpfulness seems like the right idea, coming at the right time.
How do you make sure that the system works the way you hope it will? How do you prevent it from being manipulated in any way?
We have built both programmatic, artificial intelligence and a human touch into Jelly’s core. Trust is earned on Jelly. Yet, it can be taken away just as easily. People who are signed up to Jelly have a public profile displaying Trust count, Helpful count, and the topics they are most helpful within.
Jelly is a self-policing network with people deciding who is helpful and who is trustworthy. As a failsafe, we also have a flagging system that will alert, again, both algorithms and people to act on anything that is considered out of bounds for Jelly.
You mentioned in your Medium post that as “the Follow model” -- in which social network users opt in to see the content of others -- has become more widely used, it's become more complicated. In what sense?
What I meant, in a larger sense, is that the basic concept of a recipient-driven system has been compromised.
When I originally designed the Follow model, I made a point of highlighting that the recipient is in complete control of what content they choose to see. I made a comparison to email. If someone knows your email, they can email you any time and anything they want -- you are not in control. The Follow model, as a recipient-driven system, meant this couldn’t happen. You would only see what you chose to see. Also, you could choose to not see it anymore by unfollowing.
Most of the services that employ the Follow model have since compromised this recipient-driven aspect by adding features allowing for the same thing to happen as the email example I just mentioned. If they know your @name, anyone can get a message in front of you whether you want it or not. This is not what I envisioned at the start.
Jelly users ask questions anonymously, but in many corners of the internet, anonymity can breed harmful interactions. More broadly, what do you think social platforms have to do to create safer spaces? What responsibilities do they have to community members?
Agreed. Anonymity on the web can be a serious problem in some cases and a life saver in others. Asking anonymously is key to Jelly because you are not being judged, nobody knows it was you who asked and you are free to ask questions without fear.
However, at Jelly, we take precautions to make sure the brighter side of anonymity prevails. Questions submitted to Jelly are not simply routed to others. Our question routing system is a stepped system of programmatic human-in-the-loop AI. Simply put, there are some questions that will not be routed to helpful people out of respect for their time and their humanity.
On a broader note, social platforms should think differently about the web vs. real life. These worlds are similar, but they don’t match up evenly. We should recognize these differences. We should celebrate the serendipity the web provides, yet not condone the hate it may breed. We can prevent and inspire based on thoughtful choices when designing social systems. We’ll make mistakes, but the beauty of creation on the web is that those mistakes can be fixed.