Lebanon has been witnessing an ongoing, constant war for the past three years. For the most part, journalists, reporters, activists, residents, and citizens wanting to learn what was happening did not automatically turn to mainstream TV channels. Rather, they turned to WhatsApp, Meta’s messaging app widely used in the Global South by people of all ages. News channels, whether directly affiliated with a major outlet or community-led, took over, and people started relying on them for information. Needless to say that with the absence of a proper oversight authority regulating the spread of news, mis/disinformation campaigns took over. However, the issue here extends beyond that: Meta and other similar big tech companies decided, thousands of kilometers away in California, to enforce how Arabic content was ranked or moderated with no warning and no recourse, and most importantly without the context needed to accurately assess takedowns or bans. Recent media studies scholars proved this very argument: that 59% of surveyed participants used WhatsApp for wartime news, and that the reliance on social media is now stronger than ever.
Lebanon is not an isolated case. Across South West Asia and North Africa (SWANA) and the Global Majority as a whole, the same pattern holds. Meta’s family of apps, alongside YouTube and TikTok, constitute the primary infrastructure through which hundreds of millions of people access information, follow news, and participate in public life. There is no meaningful alternative. These tech companies’ headquarters, leadership, and core priorities are located far from the communities that rely on their services the most. In one of our class presentations, we discussed that controlling information has always been key to maintaining power. What has changed now is not the intent to control, but the mechanism and the actor wielding control. Today, these companies possess personal and private data about each of their users. Lebanon’s political paralysis and wars with Israel have created an information vacuum that traditional media could not fill. WhatsApp for Lebanese people is not merely a messaging app but the primary medium to learn about what’s going on.
The reliance on Meta-owned platforms for wartime information represents a fundamental shift in the digital autonomy of users. Lebanese and Arab people are not using WhatsApp out of complete free will, they are using it because it is the only accessible infrastructure available. This is a structural vulnerability, a single American company controls the infrastructure through which a population accesses critical information during conflict, that company holds extraordinary power, power that can be exercised through content moderation, algorithmic ranking, or account suspension. Add to that the metadata Meta and WhatsApp collect, which can create highly sophisticated user profiles. When Iran shut down the internet in January 2026 in response to protests, citizens knew that the state was responsible for the loss of connectivity. Same for the war that followed. The control was explicit, and resistance could be explicitly organized. Meta, in contrast, removes content or or bans an account, or a news group chat onWhatsApp for “promoting” a dangerous individual or organization (DOI) when in reality the news required to share an image or content about this person or organization. In cases where Meta deprioritizes an account reporting on the war for example, users do not know if their posts failed to distribute due to technical error, algorithmic ranking, or explicit removal. They cannot appeal to elected representatives or invoke legal protections because Meta is a private actor, not bound by constitutional constraints on free speech. The most likely near-term development is Meta’s expansion of automated content moderation powered by LLMs.
Current moderation currently relies on a combination of human reviewers and algorithmic flagging; but future systems will likely invert this ratio, with algorithms making primary decisions and humans handling only appeals or edge cases. This shift will be justified as necessary for scale, as Meta currently moderates billions of pieces of content daily, and human review incurs more costs on the company. But automated systems will very likely amplify existing biases. LLMs trained predominantly on English-language (or Western languages) data will perform poorly on Arabic, particularly on colloquial Arabic used during crisis situations. In my work at a regional digital rights CSO, we’ve proven time and time again how important human moderators who actually understand dialects and local contexts can be. One Arabic word in Libya can mean an entirely different thing in Syria. The result will be the systematic over-removal of Arabic-language content, particularly content addressing political violence. This creates a perverse outcome. English-language discussions of the Lebanon crisis remain visible and accessible, Arabic-language accounts from actual, grassroots organizations face disproportionate removal. Information control becomes encoded into the algorithm itself, instead of visible to external observers and also difficult to contest. Alternatively, regulatory intervention may force Meta to implement different moderation standards in different regions. The EU’s DSA already mandates transparency and context-specific moderation and other regions may very well follow in their footsteps.
This could create a version of the internet where content moderation varies dramatically by jurisdiction, like a Lebanese version of Meta with different rules than the global version. We’ve already seen this happen with Elon Musk’s X under the influence of powerful Gulf governments, where certain content was banned or hidden for people falling in a certain geographical location (i.e. “geoblocking”). A third, but unrealistic scenario in my own opinion is that states can respond by nationalizing digital infrastructure, like in China. They created domestic alternatives to Western platforms (WeChat, Telegram proxies). But this is highly unlikely to happen in the Global Majority because it requires significant resources that these states do not have. In Lebanon, the state might establish a national telecommunications authority that mandates local hosting of messaging data and local moderation of content. But this doesn’t really solve the issue. It just transfers corporate control to state control. But it creates the possibility of democratic citizen oversight. If digital infrastructure is state-owned, citizens can theoretically demand accountability through political processes and public opinion. The risk is that statecontrol becomes more efficient and less transparent than corporate control. The most optimistic scenario involves a critical mass of users abandoning centralized platforms for decentralized alternatives. Signal is already functionally superior to WhatsApp for privacy. If the app achieves sufficient success, users may consider the switch. This scenario is least likely in the medium term because big tech is powerful. WhatsApp’s value comes from its ubiquity, and no alternative currently has comparable reach in the Global South.
The ideal outcome is not to eliminate information control, of course, some content moderation is necessary to prevent the proliferation of hate speech and dangerous content that could harm vulnerable people like children. The goal is to democratize and distribute control so that no single actor, corporate or governmental, controls what citizens can access. This requires coordinated action across political, regulatory, technical, and social domains. Lebanon should establish regional digital sovereignty frameworks requiring Meta to operate moderation hubs staffed by native Arabic speakers and local experts, maintain data within Lebanese territory, and allow different messaging apps to work together, so you can use any app and still reach your friends, no matter which app they use. Something like a secure SMS system. Governments must simultaneously implement rigorous regulatory accountability. This means mandating transparent moderation reports classified by region, language, and category, not Meta’s vague voluntary disclosures. Independent appeals bodies composed of government, civil society, and legal experts should have authority to overturn moderation decisions and impose fines. Platforms must undergo independent algorithmic inspections before deciding and implementing new systems, and users must have the right to know why content is removed with meaningful recourse. Locally, Lebanon or the country in question should invest in technical alternatives that reduce dependence on Meta. Open-source platforms like Signal should be deployed with government support as community-operated alternatives in which users have more control of their data. Civil society also plays an essential role in generating demand for these changes. Organizations should educate users about algorithmic ranking and content moderation, publicize Meta’s failures through transparency data, organize strategic usage reductions signaling demand for alternatives, and partner with advertisers to create economic pressure.
