I want social space online, but none of the social media software currently in existence meets my needs. Beyond that, I believe that many of the offerings are actively harmful to privacy, security, and democracy. I know many people feel the same way, but can't opt out for lack of alternatives. Clearly it's time to build something new. Something that's useful, effective, and responsible, but also attractive. What elements will it need?
The scope of this post is laying out the broad strokes of what a good social media system would need, not what it would actually look like. I plan on writing up at least two more posts: One on considerations that any social media software must address, and another on a possible architecture.
These are my top priorities for a good social media system:
- Civil atmosphere: This is a little more than "people are friendly". I want people to be approximately as civil as they are in person. Many of the things people say to each other online are horrifying, but often they would not say such things face-to-face. So, friendly even when in disagreement. Feeling safe to have disagreements, even on delicate or highly partisan topics. Structural elements of communication software can have a strong effect on this.
- Resistant to censorship: This includes both governmental and corporate censorship. Legal authorities (local or remote) and service providers (network operators, hosting operators, computer or phone manufacturers) must not be capable of removing, blocking, or suppressing content in a sweeping fashion. This includes by means of pressuring admins, if the system even has admins. (Censorship is always possible, but it should be expensive, and scale in cost linearly with people affected.) This likely precludes policing of the system for content that is illegal in one jurisdiction or another.
- Resistant to surveillance: Content that looks private must be private (unlike a "private message" on Facebook). Again, applies to both governmental and corporate surveillance. Attacks must be individualized (costs should scale at least linearly).
- Resistant to psyops (ha, spot the 2018 addition): Turns out if you expose the conversations and relationships of millions of people and allow targeted messaging to those people, you can maybe sow diviseness and hatred and swing elections one way or another. (I'm referring, of course, to Facebook and Cambridge Analytica.) The system should do what it can to avoid facilitating political warfare and other large-scale psychological influence campaigns.
- Accessible and modifiable: Users must be able to modify their interactions with the system to support their unique needs, and to shape it to their unique desires. Must support users on both desktops and phones, on bad internet connections, and behind NATs.
- Non-technical interface: If the software requires running your own server or understanding the intricacies of PGP and TLS, it's a non-starter. Related to accessibility, but deserves a top-level mention.
- Friendly to faceted identity: Users are empowered to present different faces to different people at different times. Unified online identity is a luxury of the few, and even in the best case eventually leads to unpleasant context collapse. Nolan Lawson explains the need for multi-faceted identity in the first half of this post on decentralized social networks.
- Data durability: Make it easy for users to retain their data even if a server goes down permanently. Facilitate continuous, automatic backups for users who prioritize this. (Compare to e.g. Facebook where you can get a data export, but only via a manual process.)
Nice to have
Attributes that would be nice to have, if they're feasible and do not lead to substantial compromise on the top priorities:
- Substrate for organization: Twitter and Facebook are pretty good for this, except with the strong risk of surveillance. I think it's important that something fill this role, but a system focused on social media might not be the place for it.
Attributes that I'm not sure if I want or not:
- Allows fast and free flow of ideas (and fine, memes as well): The ability to communicate globally is one of the great wonders of the internet. The software should not unduly raise walls against sharing ideas. However, I sometimes wonder if we are enabled to share too quickly, such that falsehoods are spread much faster than their corrections (and we know that corrections have little effect on perceptions.) « A lie can run round the world before the truth has got its boots on. » (Terry Pratchett, The Truth). Is there a balance to be struck?
- Fun and attractive: People are only willing to use so many social media systems at once; to some degree there's competition against Twitter and Facebook for users. But the latter are using dirty tricks for the purpose of "engagement". Fast feedback, videos, making as much public and charged as possible. I don't think a good alternative is to do the exact same thing under a different name. It has to be enjoyable, but should not be addictive. The users should get what they need out of it and then sign off and go live their lives.
You'll notice I haven't said anything about encryption, where the data would live, who runs the software, etc. Those are all questions about how to meet these goals, and I think it's important to first lay out what we want. Or at least what I want. ;-) I'll lay out my thoughts on these in a future post.
What other high-level goals am I missing?