Thursday, November 19, 2026

Subscribe

NewsEdition UgandaUganda

Uganda plans AI tool to flag online content, but not decide appeals alone

Uganda’s UCC is preparing an AI tool to detect some online content. The project raises questions about errors, appeals and data protection.

Ugandan professionals gathered in Kampala to discuss the regulation of artificial intelligence and online content

In Kampala, the regulation of digital content is entering a new phase. Uganda plans to use artificial intelligence to identify certain content before it is shared. The project is aimed in particular at sexually explicit images and videos, including those created or manipulated using generative tools.

A project led by the communications regulator

The Uganda Communications Commission (UCC), the authority responsible for overseeing communications services, is preparing to acquire a system capable of identifying certain words and content shared on digital platforms. The Minister of ICT and National Guidance, Justine Kasule Lumumba, announced the plan in a television interview on 25 September 2026.

The system is expected to use artificial intelligence to detect content deemed harmful or inappropriate. The authorities cite sexual content in particular, whose circulation is a concern for families, child-protection organisations and several players in the education sector.

The minister also referred to the growing use of images and videos generated by artificial intelligence. These tools make it possible to create realistic content quickly, sometimes without the consent of the people depicted. Its circulation can therefore raise concerns about child protection, privacy and image manipulation.

For now, Kampala has not specified the name of the supplier, the cost of the system or its deployment timetable. The authorities have also not detailed which platforms are targeted. It is therefore unclear whether the system will be able to act directly on social networks or merely send alerts to the regulator.

Regulation already in place, but little automation so far

The initiative is part of an existing regulatory framework. Ugandan rules adopted in 2019 cover audiovisual content, child protection and the distribution of sexually explicit content, among other areas. They provide for restrictions on realistic depictions of sexual acts or nudity, while distinguishing educational and health content shared responsibly.

The use of AI would therefore not necessarily create a new ban. It would mainly change how the rules are monitored and enforced. Until now, oversight has relied largely on reports, human checks and action by platforms. An automated tool could examine a far larger volume of posts.

However, this development carries a key risk: a machine may mistake illegal content for a journalistic, medical, artistic or educational post. A photograph intended to inform people about sexual health is not subject to the same rules as a video exploiting a child. The system’s effectiveness will therefore depend on its criteria, as well as on whether decisions can be reviewed.

Language support is another sensitive issue. English and Swahili are among Uganda’s official languages, but daily life also involves several national languages. A system trained mainly on English-language data may struggle to understand local expressions, metaphors and the ways Ugandan communities communicate.

Safeguards remain an open question

The project comes as the country is still developing its broader framework for AI governance. Since 2026, Uganda’s Ministry of ICT has been holding consultations to draw up a National Strategy on AI and Emerging Technologies. The document is intended to guide the use of AI in public administration, agriculture, health, education and digital services.

In March 2026, the ministry said the consultations would cover governance, ethics, standards, infrastructure and investment. The government aims to support innovation while addressing risks linked to data, cybersecurity and automated decisions.

In May, representatives of government, universities, the private sector and civil society also reviewed an assessment of Uganda’s AI readiness, conducted with UNESCO. The document identifies progress in research and innovation, but also gaps in specific regulation, data governance, infrastructure and skills.

These efforts make the UCC’s planned tool particularly significant. A content-filtering technology should not operate outside a clear framework. Users need to know what content is being analysed, under which rules and with what safeguards. Platforms must also be able to explain decisions made on the basis of an automated flag.

Data protection is another concern. Uganda adopted a Data Protection and Privacy Act in 2019. However, the assessment conducted with UNESCO notes that existing rules do not always cover issues specific to AI, such as algorithmic transparency, automated decisions or the representativeness of the data used to train systems.

A debate that extends beyond Uganda’s borders

The issue directly affects East Africa. Uganda is a member of the East African Community, or EAC, a regional organisation working to strengthen economic integration and co-operation among its member states. Digital platforms operate across the region, while content hosted in one country can be accessed from several others.

A strictly national approach could therefore create different rules for users connected to the same services. Conversely, Uganda’s experience could inform regional discussions on child protection, AI-generated content and platform responsibility.

For the authorities, the challenge will be to protect children without turning automation into an opaque censorship mechanism. For users, the question will be very concrete: what happens when a post is wrongly blocked? For technology companies, the debate will focus on co-operation with the regulator, data retention and the speed of appeals.

The next step will therefore be less technical than legal and institutional. Kampala will need to specify the categories of content targeted, the platforms concerned, the role of human agents and the mechanisms for challenging decisions. The national AI strategy, along with discussions within the EAC and the African Union, will show whether Uganda favours automated surveillance or regulation backed by verifiable safeguards.

Automatic translation. Original French version

The quiz

Edition UgandaUganda in five questions

5 questions