Africa's tech policy community is converging on one question that has no consensus answer: who bears legal responsibility when online platforms host harmful content, and how do governments enforce accountability without becoming censors themselves?
Three policy developments in the past eighteen months set the parameters of this debate. South Africa's Competition Commission released its long-awaited Media and Digital Platforms Market Inquiry (MDPMI) final report on 13 November 2025, proposing both structural remedies and content moderation regulations. The African Commission on Human and Peoples' Rights adopted Resolution 630/2025 on 11 March 2025, mandating the development of continental guidelines for platform monitoring. And Kenya's Computer Misuse and Cybercrimes (Amendment) Act 2025 handed regulators new power to order websites inaccessible when they "promote unlawful activities." The contrast between these approaches tells us almost everything we need to know about what works and what does not.
The Case for Platform Accountability
The advocates pressing for stronger intermediary accountability are not wrong to be concerned. The African Commission's Resolution 630/2025 articulates a real problem: major platforms including Meta and YouTube have quietly replaced professional, multilingual fact-checking operations with crowd-sourced Community Notes systems that struggle with African-language content. The Commission correctly notes that Community Notes are "susceptible to be captured by forces that do not respect human rights," and are inadequate substitutes for independent editorial review during elections, health crises, and armed conflicts.
The stakes are concrete. A study by Internet Sans Frontières and the KenSafeSpace coalition covering January through July 2025 found that 43% of analyzed Kenyan social media content showed indicators of hate speech, predominantly ethnic and religious in character. In Ethiopia and Sudan, civil society organizations have documented how hate speech propagated through under-moderated platforms has contributed to violence. When platforms retreat from moderation and African-language content receives minimal scrutiny, the costs are borne by African users—not by platform executives in Menlo Park.
South Africa's Structural Approach
South Africa's MDPMI final report offers the most sophisticated African answer to this problem—and it is notably not primarily a liability framework. Rather than imposing takedown mandates or criminal penalties on platforms that fail to remove content, the Competition Commission focused on structural power imbalances. The Commission found that Google, Meta, Microsoft, TikTok, X, and AI companies collectively dominate the gateways through which South Africans access information, while paying inadequate compensation for the local news content that makes their services valuable.
The remedies are proportionate and market-oriented. Google and YouTube agreed to pay approximately R688 million (around $42 million) in content licensing, innovation grants, and capacity-building for South African media. Meta must establish a local liaison office, remove follower thresholds for monetization, and provide training and advertising credits. The Commission further recommended that the Department of Communication and Digital Technology develop content moderation regulations under the Electronic Communications and Transactions Act (ECTA), and establish an independent Social Media Ombud to handle public complaints.
The MDPMI demonstrates that competition law can drive substantial platform accountability without requiring governments to build content removal regimes that risk weaponization against critics.
The Social Media Ombud proposal deserves particular attention. Unlike notice-and-takedown systems that put platforms in the position of adjudicating legality under political pressure, an independent ombud creates a structural check—a body with standing to investigate systemic moderation failures, including gaps in African-language coverage, without creating individual post-level censorship powers.
The Censorship Risk Hiding in Plain Sight
The danger lurking in Africa's platform accountability moment is that governments adopt the vocabulary of intermediary responsibility while building tools designed for political control. Kenya's Computer Misuse and Cybercrimes (Amendment) Act 2025 is a case study. Section 6 of the amended Act empowers regulators to make entire websites or applications "inaccessible" when deemed to "promote unlawful activities"—a standard broad enough to cover opposition political content, journalism about government misconduct, or protest organizing.
This is not hypothetical. Kenya's government ordered internet disruptions during anti-government protests in 2024. Freedom House rated Kenya "partly free" in its Freedom of the Net 2025 report. Adding blanket website-blocking authority to a government with that track record does not produce platform accountability—it produces a censorship instrument dressed in the language of content moderation.
Nigeria's NITDA Code of Practice for Interactive Computer Service Platforms (September 2022) takes a more measured approach, establishing a conditional safe harbor: platforms face no liability for third-party content if they act on valid takedown notices. This broadly tracks the EU's e-Commerce Directive model and offers a workable baseline. But even the NITDA Code requires large platforms to maintain local representatives in Nigeria, creating compliance chokepoints that regulators could exploit to pressure editorial decisions.
What a Proportionate Framework Needs
The African Commission's ongoing guidelines process—launched at the Digital Rights and Inclusion Forum in Lusaka in May 2025 and targeting consultations through November 2025—is the right venue to establish continental standards. Three principles should anchor any framework that emerges.
First, no proactive monitoring obligation. The African Declaration of Principles on Freedom of Expression and Access to Information (2019) is explicit: states should not require intermediaries to monitor content they did not author. Any ECTA-style content moderation regulation must preserve this baseline; departing from it collapses the distinction between accountability and surveillance.
Second, human rights impact assessments before policy changes. Resolution 630/2025 mandates that platforms conduct transparent impact assessments before changes that could affect elections or conflict zones. This is the kind of due diligence requirement that imposes real accountability without general liability for third-party content.
Third, structural remedies over criminal liability. South Africa's approach—revenue sharing, algorithmic fairness, independent ombuds, African-language training obligations—is a more durable solution than penalties for content failures. Criminal liability for hosted content creates incentives for over-removal that silences legitimate speech; market-based remedies align platform incentives with user welfare.
Africa's platform accountability debate is maturing. The question is no longer whether to hold platforms to standards—but which instruments do that without handing governments a key to every digital lock on the continent.