Twitter has accepted a buyout offer from Tesla and SpaceX billionaire Elon Musk, leading to a day of frenzied speculation over one question: how is Musk going to change Twitter?
Elon Musk’s Twitter plans are a huge can of worms
Online speech is a minefield, and if Musk really intends on a minimally moderated Twitter worldwide, he could expect huge fights in countries that restrict things like hate speech and false information. But Musk’s view of free speech doesn’t seem too concerned with that. In a TED interview, he indicated that Twitter should “match the laws of the country,” which suggests he could continue practices like region-locking certain content and follow rules like India’s social media regulations.
There are some excellent overviews of how Musk might decide to tweak Twitter’s policies and the hazards he’d face, including from The Atlantic’s Charlie Warzel and TechDirt’s Mike Masnick. But, at this point, we don’t know very much about how Musk would concretely change Twitter’s speech policies. He’d probably urge moderators to hand down fewer bans and potentially leave questionable content up. But virtually every site that claims a “free speech” banner ends up banning something that makes it deeply unpleasant for users, advertisers, or the site owners themselves — so it’s premature to say how far his commitment will go.
“Open source” algorithms
One of Musk’s areas of concern is recommendation algorithms that amplify or downrank tweets and accounts in potentially biased ways. He’s proposed publishing Twitter’s algorithmic sorting systems on Github for people to publicly review and comment on, making something like the “top tweets” ranking system more theoretically legible.
Musk has described making the algorithm “open source,” but he hasn’t outlined specific plans to follow the requirements of an open source license, so he could mean it in a more informal sense. He could also be describing something that works within Twitter’s central product or through the separate but Twitter-funded open source Bluesky project — which would have different implications for Twitter’s core app.
Spam and scam bots
Musk has indicated that “spam and scam bots” and “bot armies” are Twitter’s new Public Enemy No. 1. That makes sense, as Musk is a perennial subject of scammy crypto impersonators. How he’d police this, however, is an open question. Unlike with speech maximalism, there’s no huge philosophical difference here — nobody likes spambots! Twitter already purges fake accounts and has banned certain features, like tweeting simultaneously from multiple accounts, that facilitate bot spam. So how would Musk do better?
Well, Musk could have some kind of hitherto-unannounced anti-spam tool in the works, although there’s no indication he’s spent more time thinking about this than Twitter’s own engineers have. (Again: Twitter has lots of incentives to police spam already!) Or Musk could simply decide to err far more heavily on blocking non-malicious automated account activity, locking down access to Twitter’s API, or demoting content from humans who act too much like bots.
Unfortunately, that goal would probably work in conflict with his push for freedom of expression and transparency. As mentioned above, publishing the inner workings of Twitter’s amplification system would also give spammers more tools to work with. And a strict automation crackdown could block bots that perform interesting and valuable services on Twitter.
Authenticate all humans?
The weirdest and arguably most disruptive part of Musk’s Twitter pitch lies in his last three words: “authenticating all humans.” Musk made a similar comment on Twitter before the purchase, phrasing it as “authenticate all real humans,” following a commitment to defeat bots. He hasn’t been specific about the goal of this authentication, though, or how it would be carried out.
“Authentication” could potentially mean a couple of different things here. It could refer to people having to pass some kind of captcha-style “am I a human” test to post — although, as with spambot bans, if there were an easy way to do this without affecting good-faith users, Twitter would probably have done it already. It could also mean asking people to submit identification that proves they’re specific humans, either to receive a verification checkmark (something Musk has previously suggested) or to operate on the service at all.
Twitter has a long-standing commitment to allowing anonymous or pseudonymous speech, even submitting legal briefs arguing for its benefits. Asking users to de-anonymize themselves undercuts that commitment significantly. Even if a name isn’t revealed to other users, collecting information on real identities offers a trove of information for governments to request, and it’s vulnerable to hacks or security flaws. “There are no easy ways to require verification without wreaking havoc for some users, and for free speech,” the Electronic Frontier Foundation noted yesterday.