Since completing his acquisition of Twitter, Elon Musk has made changes to bring back banned members, allow more types of content, and in general reduce moderation.
Recently, for example, Musk decided that Twitter would stop policing Covid-19 misinformation. After an informal poll showed overwhelming support for restoring banned accounts, Twitter is now returning privileges to 62,000 banned members.
It’s not clear how moderation staff was affected by the combination of a 50% company-wide headcount reduction, cuts in contractor staff, and unforced resignations. But, the effect has likely been significant. High-profile departures include Yoel Roth, Twitter’s former Head of Trust & Safety.
Bots and Spam
Musk has promised to eliminate, or at least drastically reduce, “bot” accounts and spam posts. This would be a significant and welcome achievement. No successful community can be allow unchecked automated posting or a deluge of low quality, spammy content. Having to wade through irrelevant or even deceptive posts degrades user experience.
Previous Twitter management downplayed the number of bot users while Musk estimated that as much as twenty percent of Twitter’s user base might be fake. One cybersecurity firm calculated a value of at least 12%. Whatever the truth, ensuring that Twitter members are real people and that nobody, bot or human, is posting spam, is a praiseworthy goal.
Musk may be making headway in this effort, but how much isn’t clear. Yesterday, CNN reported that searches for information on protests in China were deluged with spam and porn posts. The persistence of the spam content has been blamed on a reduction in human moderation and a reliance on automated tools.
How Free Should Speech Be?
Free speech is a bedrock of democracy and seems like a laudable goal for a large online community like Twitter. Unfortunately, the reality of unfettered speech online is quite different than the idealized version. This is particularly true when members can post anonymously, as many do on Twitter.
I’ve moderated and built online communities since the days of Compuserve, and in my experience the most successful, thriving communities all have clear rules and effective moderation. Most normal people don’t want to be shouted at or insulted. If they have a question, they don’t want to be fed false information.
New arrivals to a community are particularly vulnerable. If their first experience is negative – they are treated rudely, for example – they may never return. Even community veterans will leave if they are subject to constant negativity.
Unmoderated speech online inevitably devolves into arguments and discourtesy. Even in a community where most members behave in a responsible way, a small percentage of disruptive members can poison the experience.
Communities that exhibit growth, stickiness, and high levels of member participation cultivate a welcoming and helpful environment.
Twitter vs. Other Communities
Musk faces a serious challenge, perhaps many serious challenges.
Successful communities tend to organize around topics of common interest. FlyerTalk.com, for example, focuses on air travel in general and frequent flyers in particular, generating millions of visits per month. Topical communities like these can simply ban discussion of third-rail issues like politics, race, and religion. They can also set tight rules for content relevance, profanity, courtesy to other members, and so on.
Even a large, general community like Reddit is divided into tens of thousands of subreddits, each of which can establish rules and moderation practices. Similarly, Facebook groups offer a controlled, moderated experience where disruptive individuals can be removed.
The overall Facebook experience is more controlled than Twitter. Interactions tend to be with a smaller group of accounts, most of them known to the member. Accounts can’t be anonymous, although that doesn’t stop spammers. Facebook data shows they remove a billion and a half fake accounts every quarter.
In contrast, Twitter is, in essence, one giant group of members. Creating rules of engagement is difficult and limiting topics is nearly impossible. Political discussion is the reason many members visit Twitter. Memes that annoy some members are happily devoured and shared by others. Insults directed at both public figures and other members are part of the experience.
Can AI moderation eventually do as good a job as humans? Or even better? Perhaps someday. Either way, establishing exactly what kind of speech is allowed is still a human decision made by Twitter’s leadership.
Could Twitter create walled gardens, a bit like Reddit, where groups of members could establish rules and create their own environment? Perhaps, but I’ve seen no inclination to do this.
A Platform for Everybody…
When anyone can join a community without fear of being silenced or removed, many members will be emboldened to say anything… inaccurate information, rude language, intolerance of others’ views, and more.
When anyone can say anything, discussion will devolve to the lowest common denominator. Those members who are seeking a thoughtful interchange of ideas or merely a pleasant social experience will ultimately leave the community. Members that remain will yell at each other, never changing anyone’s mind but creating friction and unpleasantness.
Musk has emphasized the importance of revenue growth for Twitter’s survival. A community filled with animosity will attract neither advertisers nor members willing to pay to participate. If this is the direction Twitter is headed in, the community and the business itself are in trouble.
A platform for everybody is ultimately a platform for nobody.