There are three areas — content moderation, transparency, and data access — to watch closely as Musk takes ownership of Twitter.
Area of Study
This article was originally published at Tech Policy Press.
With Twitter agreeing to Elon Musk’s offer to take the company private, the wealthiest person in the world will soon own the global social network that is perhaps the most popular among politicians, journalists, activists, academics, and others. While the deal needs to be finalized, focus on whether Musk would (or could) complete the acquisition has now shifted to how he will change the platform.
Musk has signaled an interest in making a number of changes to Twitter, and prognostications are circulating about what Musk will actually do once the deal is completed. Given that the one certainty with Musk is uncertainty, our goal is not to offer predictions about how the company will change under new ownership, but instead to suggest the value of focusing on three key areas for monitoring when (if?) the transfer of ownership is eventually completed.
In a press release marking the deal, Musk foregrounded speech concerns: “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.”
It is, therefore, not really going out on a limb to suggest that Musk is likely to focus on content moderation (i.e. labeling and removing content that runs afoul of rules, such as hate speech and misinformation) and deplatforming (i.e. temporarily or permanently suspending accounts that frequently break platform rules). These actions are the product of complex systems that include automated classification methods, human reviewers, and professional fact-checkers. And the platform — even without changes in ownership — is constantly adjusting how this system operates. In recent years, Twitter has invested heavily in specialized teams focused on addressing core potential harms on the platform, such as removing hate speech, limiting the spread of misinformation, and reducing spam.
But as Techdirt’s Mike Masnick noted, Musk’s vision for online speech does not seem to recognize this reality. Instead, his views of social media align with those of tech founders in the early 2000s: a vision of a town square where everyone can be heard, and good speech will counter bad. But this vision didn’t come to pass, especially at scale. Platforms have spent the past two decades developing both the ethics and practice of trust and safety. While far from perfect, Twitter has made strides in key areas, and Musk could undo or de-resource this work in the name of his free speech vision.
Musk has also signaled an interest in increasing transparency around the way Twitter operates, which, ironically, will also be key to measuring Musk’s impact on Twitter in at least two ways.
The first is around content moderation. We can observe this through the platform’s policies, which define allowable content and describe the platform’s actions for breaking its rules. And we can observe this in the platform’s enforcement of those policies — in other words, the actual actions it takes on accounts and tweets. While the former is easier to observe through company blog posts, the latter is far more influential but often opaque and difficult to measure at Twitter’s scale (and thus will require substantial platform data).
The second is around the product itself. Musk has already signaled support for several product changes, ranging from small updates to give users more control (adding an edit button, allowing long-form tweets) to larger changes to increase trust in the platform (getting rid of spam bots, authenticating “all real humans”). Most notable, however, is his promise of full algorithmic transparency, which would allow the public to see the complicated calculus deciding what appears on our feeds. In addition to providing users with more information about their experience on the platform, it would also enable academics to better monitor how Twitter changes its algorithm over time. Whether Musk the owner will feel the same about this position once it is his proprietary intellectual property that he is proposing to open source is another matter, but should he actually improve algorithmic transparency, it would be a welcome development.
(As an aside, authenticating “all real humans” could end up making Twitter a much less useful tool for pro-democracy activists in authoritarian regimes if it ends up removing the ability of people to tweet anonymously. While the removal of anonymous accounts might be welcomed by many who have been the target of attacks from such accounts, it is important to remember the importance of anonymity in countries where opposition activists face threats to physical safety and/or imprisonment.)
Vocal groups of skeptics and optimists have already emerged in response to Musk’s acquisition. A key challenge will be that social media data are massive and optimized for search, meaning that it will be easy to find evidence to substantiate any expected change. The platform could become a Rorschach test of sorts for those expecting Twitter to enable more misinformation and hate speech and those expecting it to become more aligned with “free speech.” Rather than relying on conventional wisdom and anecdata, rigorous research will be needed to understand the “Musk effect” on Twitter.
Relative to other major social media platforms, Twitter data has been the most accessible to outside researchers. As a result, a majority of quantitative research on social media and politics has focused on Twitter, enabling a relatively deep understanding of the platform.
While Musk is publicly committed to transparency, the extent to which he supports data access for external researchers is currently uncertain. From the way that academics reacted to news of Musk’s purchase on Twitter, it’s clear there will be scholarly interest in studying Musk’s impact on the platform, which will require data.
Musk has a long history of blocking Twitter users whose actions he dislikes. If academic studies paint a negative image of the social network under his ownership, there would be no easier way to block this work than cutting off data access.
This is precisely why we need federal legislation to mandate external access to social media data. Musk will be the second billionaire solely in charge of a major social network. In an age when social media and the internet have reduced our reliance on traditional media gatekeepers, Musk and Meta’s Mark Zuckerberg have become a new type of digital gatekeeper. Indeed, they hold the keys to the very data that would enable external researchers to study the platforms themselves. Twitter’s relative openness to external researchers was entirely at the whim of platform leadership. The Musk acquisition shows how quickly the status quo regarding leadership can change – and therefore the need for regulation to ensure that the ability to conduct research on the impact of these platforms on society is not left up to the whims of their owners.