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? Musk’s stated plans are a series of features and principles that he may or may not be serious about pursuing, but they demonstrate a potentially conflicting series of goals and changes whose mechanics have been outlined very little — if at all.
Musk named his priorities in a press release, echoing earlier statements he’s made about potential changes. “Free speech is the bedrock of a functioning democracy, and Twitter is the digital town square where matters vital to the future of humanity are debated,” he said. “I also want to make Twitter better than ever by enhancing the product with new features, making the algorithms open source to increase trust, defeating the spam bots, and authenticating all humans.”
In other words, he’s got four main ideas for unlocking Twitter’s potential, and every single one is a huge can of worms. Let’s break them down one at a time.
Free speech
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 following rules like India’s social media regulations.
Musk has a lot more leeway over changing Twitter’s policies around what kinds of content are banned, of course, and when users are suspended. He’s indicated that he would rather err on the side of “time-outs” and leave borderline content online. Many have speculated that this would return former President Donald Trump to the platform, which isn’t an unreasonable prediction, but Musk hasn’t said anything about it publicly. (Trump has also claimed he wouldn’t come back.)
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 would err on the side of leaving 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.
Transparency is generally welcome, and Twitter’s former CEO Jack Dorsey has also suggested letting users pick between different recommendation systems. That said, many web platforms (including Google and Reddit) don’t disclose precisely how their systems work because that would give spammers and other bad actors a guidebook for gaming the system. Twitter’s algorithm also won’t explain how any given tweet was prioritized unless Twitter releases a huge amount of supplementary data, nor would it necessarily illuminate the rationale behind any human moderation that intersects with it. And it would be incredibly vulnerable to people who want to make bad-faith claims by taking pieces of it out of context, willfully misinterpreting them, or sowing conspiracy theories about them.
Beyond that, Musk hasn’t described how he’d integrate any suggestions made by other developers or readers — which, again, would probably involve a lot of spammers — into Twitter’s algorithm. Maybe he could follow the path suggested by Dorsey and let people fork their own versions of Twitter’s recommendation system, turning it into an actual open-source system? Maybe he could set up a Facebook Oversight Board-style committee that would approve suggested changes? We won’t know for a while.
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 — like Big Tech Alert, which tracks who Silicon Valley’s big players (including Musk) are following and unfollowing, or Editing TheGrayLady, which illuminates how The New York Times tweaks its headlines and copy over time. Bots are a long-standing and beloved part of Twitter, and separating a good bot from a bad bot might be harder than Musk thinks.
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.
Musk likes to throw out odd ideas as a provocation, so his statements yesterday may not end up reflecting where the platform goes. If Twitter’s past moderation challenges are any indication, though, every change is going to open up a whole new set of questions to answer. The open question is how much interest Musk has in managing the fallout.
wazup.