This blog post was authored by Catherine Powell, adjunct senior fellow at the Council on Foreign Relations in the Women and Foreign Policy program and the Digital and Cyberspace Policy program. Additionally, she is a visiting scholar at the Center on Human Rights and Global Justice at NYU School of Law, where she also serves as a non-resident fellow at the Reiss Center on Law & Security.
On April 25, news broke that Elon Musk and Twitter had reached a deal wherein Musk would buy Twitter for $44 billion. When and whether the deal will actually be finalized is up in the air, however. Musk’s recent tweet that the Twitter deal is on hold, combined with his call for the U.S. Security and Exchange Commission (SEC) to investigate the amount of spam and bot accounts on Twitter, has contributed to uncertainty among investors about the future of the deal. Twitter stock has dropped, and the departure of high-level staff at Twitter has also signaled that the future of Musk’s Twitter acquisition is murky.
Along with uncertainty over Musk’s Twitter acquisition, his provocative criticism of the platform’s content moderation policies—concerning disinformation, hate speech, and harassment—has also sparked debate about the meaning and importance of free speech in digital spaces. As PEN America CEO Suzanne Nossel notes, it’s as if Musk has tried to take humans out of the loop of driving with his self-driving cars, but it is not so easy to take humans out of the loop when considering the impacts of disinformation or other speech considered harmful. As Nossel recently tweeted, “Elon Musk will learn the same lesson about self-governing social media as he has about self-driving vehicles, which is that they unavoidably crash.” As I’ve discussed elsewhere, pioneering scholars, such as Sarah Roberts, have documented the important role of “invisible workers” along the global digital assembly line in untangling the complexities of content moderation.
Much of the debate about the Musk acquisition fails to appreciate Twitter’s international reach and the applicability of international standards, including those governing speech. Twitter and other social media platforms operate in a number of countries whose governments have routinely restricted speech—sometimes based on “benign” reasons (such as preventing violence), but other times to suppress political opponents and critics.
Last month, I moderated a panel called “Can You Hear Me? Speech and Power in the Global Digital Town Square,” at the American Society of International Law (ASIL) Annual Meeting in Washington, DC. Panelists (in order of speaker) included:
- Matt Perault, director of the Center on Tech Policy at UNC-Chapel Hill and former director of public policy at Facebook
- Jacquelene Mwangi, doctoral candidate at Harvard Law School and former research consultant for the Center for Intellectual Property and Information Technology Law
- Emma J Llansó, director of the Free Expression Project at the Center for Democracy and Technology
- Arsalan Suleman, counsel on international litigation and arbitration practice at Foley Hoag and former acting special envoy to the Organization of Islamic Cooperation (OIC) in the U.S. State Department
The panelists discussed a number of timely topics related to free speech on social media platforms in our global “marketplace of ideas.” One of the main themes we discussed was the tension that exists between tech companies’ stated commitment to free speech on social media platforms and laws concerning free speech that differ from country to country. Mwangi noted that normally, when U.S.-based tech companies conduct their operations in other countries without setting up local offices, U.S. laws apply to the companies’ operations. According to Mwangi, the ability of U.S.-based tech companies to operate internationally without abiding by local laws has incited backlash—including by powerful interests—leading to more countries, such as Russia, to establish local presence laws. The emergence of these laws, Llansó explained, points to a growing trend in which tech companies are required to establish a local office in-country, and therefore be subject to the country’s laws. Depending on the country, establishing a local office could have significant implications for human rights, user privacy, and content moderation, Llansó pointed out. She warned that governments may want to use local presence laws as a tool to censor free speech or sidestep companies’ privacy policies to access user data.
The adoption of local presence laws is but one example of how countries and tech companies sometimes clash over what legal norms should govern the digital space. Another instance that Llansó raised is the Russian Smart Voting App, which was run by supporters of prominent opposition candidate Aleksei Navalny. The Russian government declared the app illegal, and authorities began pressuring tech companies to remove the app from their platforms, going so far as to threaten local staff with prosecution. Eventually, Apple and Google removed the app from their app stores several days before the 2021 elections due to concerns of the safety of local staff.
Part of the problem, Perault said, is that there are not clear international norms guiding tech policies’ behavior when it comes to content moderation. While international human rights law permits and even encourages governments to ban “hate speech,” U.S. courts take a more lenient approach. For example, U.S. courts have ruled in favor of allowing Nazis to march in the predominantly Jewish neighborhood of Skokie, Illinois, and the U.S. Supreme Court struck down a law that would have prohibited burning a cross on a Black family’s lawn—in both instances, finding protected speech interests. Yet in other countries, such as Rwanda, where hateful, targeted messages on the radio facilitated the 1994 Rwandan genocide, application of international standards allowing for some restrictions on hate speech have been viewed as not appropriate, but necessary. Similarly, EU countries, such as Germany, take a more restrictive approach to “harmful” speech, for example, prohibiting Holocaust denial, due to the European experience with the atrocities of the Holocaust.
Suleman reminded us about the Myanmar government’s use of Facebook to incite violence against Rohingya Muslims and the importance of platforms using content moderation where hateful speech can lead to violence, death, and even the mass slaughter of civilians. Facebook refused to release to the government of Gambia the data of government-controlled accounts in Myanmar that had violated Facebook’s terms of service by engaging in coordinated inauthentic behavior. Gambia had requested this data to support its case against Myanmar at the International Court of Justice. The government of Gambia then sued Facebook to release the records and received a fully favorable initial order from a magistrate judge to release both public and private content from the accounts. Facebook objected to part of the magistrate judge’s order that required the release of private messages in those government-controlled accounts, and a U.S. District Court judge sided with Facebook on that specific issue.
The lack of clarity surrounding international standards for tech companies’ content moderation—and how these international norms interact with domestic approaches—has been magnified with Russia’s invasion of Ukraine. Without clear international norms, tech companies are individually left to question whether further engagement in Russia is more harmful than withdrawing their services entirely. Perault noted that while companies are clearly uncomfortable operating in Russia during the war, there is not a clear answer to this question. In the absence of specific international norms, tech companies are crafting their policies toward Russia on an ad hoc basis.
Emerging norms may establish clearer standards. Having reached a deal concerning the landmark Digital Services Act, the European Union (EU) will require companies to establish new policies and procedures to more forcefully police their platforms and remove suspect material, such as hate speech, terrorist propaganda, and other content defined as illegal by EU countries. As Llansó indicated, the Digital Services Act will require companies to regulate their algorithms and create risk assessments, among other regulations. Notably, the law will enable regulators to impose heavy fines on tech companies who do not comply with the law’s provisions. While the response from tech companies has been muted, it is possible that greater regulation of large tech companies may lead to a more coherent response to human rights abuses.
In sum, social media has been essential for movements ranging from the 2010-2011 Arab uprisings to #MeToo to #BlackLivesMatter, as discussed further in my recent post on “hashtag activism.” However, as Johnathan Haidt illustrates in his recent piece in the Atlantic, social media has not only brought people together, but it has also created divisions and even led to violence. The recent shooting in Buffalo demonstrates how violent extremists are influenced by online hate. Not only do hate groups creep from the hidden corners of bulletin boards to more mainstream websites, but hate and division are more likely to be amplified online.
A final dilemma was highlighted on our panel by Mwangi, who criticized the fact that the current discourse on speech and internet regulation tends to ignore parts of the world where the right to free speech is “dependent on the whims of state power.” Mwangi said that international law needs to address the prevalence of internet shutdowns and online censorship. Despite the threat of censorship and oppression, Mwangi highlighted the vibrancy of digital movements in Africa such as #EndSARS, #ZimbabweanLivesMatter, and #SomeoneTellCNN. While some governments have managed to successfully use social media platforms to cement their power, there are brave citizens who regularly speak out for their human rights, both online and offline. How tech companies decide to confront or acquiesce to state power remains to be seen.
You can watch the full panel here (scroll down to panel #15, April 8, 9:00am). Currently, only ASIL members have access, but the video will be made available to the public in a few weeks.