Technical and Strategic Requirements for Actually ProSocial Social Media
On the largest scale, we don’t have social media, we have antisocial media. The evidence abounds that social media is amplifying the worst of us rather than the best of us, even though most of us as individuals are trying to behave responsibly and kindly.
Part of the problem is deliberate abuse by malicious actors. Can we build a better social media? One which has immunity to bullshit, suppresses trolls, and encourages mutual respect between tribes rather than mutual hostility?
About this article
This is a living document, until such time as I say it is frozen. I wanted to get the ideas collected in one place.
It’s brainstorming for an open source project to replace social media with a distributed blog/comment engine which encourages and promotes sensible, intelligent, informed and interesting contributions. How these pieces contribute to that goal will be elaborated as this document expands.
- I am in principle still a techno-utopian. I think the internet can help us think collectively.
- But we must learn to resist organized, funded bullshit that plays on people’s paranoid impulses that emerged lately, especially on FB. 2/n
- Opposition to organized rumormongering bullshit is very weak, left right and center. Globally. It leaves an opning for profound malice. 3/n
- Libertarian-leaning web culture has always been about freedom, allowing anything, with assumption that the good stuff would rise to top. 4/n
- The “new left” allied w hippie culture always had a soft spot for heartfelt outrageous claims & distaste for reasoned counterargument 5/n
- And of course, Facebook is in the business of selling clicks, and isn’t motivated to slow the rumormongering down very much 6/n
- To rescue the idea of the internet as a tool to make us collectively smarter rather than dumber, we need mechanisms to resist bullshit 7/n
- Very difficult to separate lies from good faith errors, so it’s infeasible to enforce making lying illegal. But it must be discouraged. 8/n
- W/o detracting from principle of free speech we must therefore start thinking about how to reward responsible speech & demotivate lies 9/n
- As the right and left find this anathema and the center finds it irrelevant, support for building a lie-averse web is difficult 10/n
- It’s by no means an easy problem, especially given emerging evidence that much of the BS is centrally planned, funded and promoted 11/n
- To offer hope, I think it remains conceivable that a distributed non-profit social media network can emerge. But that’s not enough 12/n
- The new social medium must structurally tend to reward truth and suppress nonsense, yet be appealing to casual users. Hard problem! 13/n
- Conventional journalism is no longer suited for purpose, but replacing it with organized malicious rumormongering is far worse. 14/n
- For years I’ve been trying to start conversation on how we can do better social media, w/o getting critical mass. 15/n
- Recent events show a network-mediated social medium for responsible public discourse is no longer a “would be nice” but a necessity. 16/n
- If anyone is interested in helping consider how prosocial media would work I’d love to hear from you. Let’s develop specifications. 17/17
Collective cognition requires trust networks. The specs below don’t fully capture how this would work. So there are some crucial specs missing! That’s probably the part we have to think hardest about.
The keys to the kingdom are that it should be non-profit, distributed, support reputation-building connected to actual expertise, trust-building, and transparency, all without being attractive and not confusing for the end user.
0) Distributed. No central server, no single point of influence or failure, runs on RSS or similar mechnaism
1) Asymmetric follow (Twitter style) and symmetric follow (Facebook style) both supported.
2) Reactions (likes, etc.) are supported but in a way that doesn’t expose individual reactions to third parties.
3) Searchable and permanent
4) Crowdsourced but reputation-based vetting of factual assertions (this is the hard part)
5) maximum interoperability with extant networks
6) Default is to show everything you follow (either symmetrically or asymmetrically) and nothing else. Any other display algorithm is under user control.
7) Sock puppetry discouraged. Ideally user identity should be unique. I’m not sure how to achieve this.
8) Beginning user is not faced with counterintuitive abstractions (like G+ circles or Diaspora aspects)
9) Something (or things) is more fun to do on this network than on existing networks, and that (those) thing(s) support the larger goals of meaningful democratic discourse.
10) Loose specifications, allowing for a variety of micro-implementations. Let the best model or models win.
Related projects to note:
Participants in this discussion are encouraged to examine prosocial mechanisms in these projects.
consensus Algorithms https://www.youtube.com/watch?v=sE7998qfjgk
As We May Think (1945)