As Uganda prepares for its general election scheduled for January 15, 2026, digital misinformation has become a growing concern for journalists, policymakers, and election observers.
While misinformation has featured in previous Ugandan elections, the increasing availability of artificial intelligence (AI) tools has changed how political content is created, modified, and distributed.
During the early campaign period, fabricated blog posts, altered images, and manipulated videos circulated across social media platforms.
Some of these were later examined by fact-checking initiatives such as DisMis by Next Media, Obijjawa by the Media Challenge Initiative (MCI), and organisations including Debunk Initiative.
Media analysts note that even when false claims are corrected, the clarifications often reach fewer people than the original content.
Political misinformation is not new to Uganda. In the 2021 elections and earlier polls, misleading material often appeared as edited photographs, quotes falsely attributed to candidates, partisan rumours on WhatsApp, and unverified blog posts.
Although these forms of misinformation were impactful, they were usually produced using basic editing tools and were easier to identify through manual verification.
The introduction of AI tools has complicated this environment. Deepfake video generators, synthetic voice tools, and AI-assisted image manipulation software allow users to produce realistic-looking political content quickly and at low cost.
These technologies have lowered the barrier to creating visuals that appear authentic, even when they are misleading.
In mid-2025, a video that appeared to show First Lady Janet Museveni urging the public not to vote for her husband circulated widely online before being identified by fact-checkers as a deepfake.
Although the video was eventually debunked, it had already influenced online discussions, demonstrating how AI-generated content can shape political narratives before verification takes place.
Beyond individual figures, manipulated visuals have also been used to misrepresent political rally attendance.
During recent campaign activities, images and videos circulated online purporting to show massive crowds at certain rallies, while others claimed to show empty venues for rival candidates.
Media analysts note that some of these visuals were either taken out of context, digitally altered, or generated using AI tools that replicate crowd scenes.
Such content can give audiences a distorted impression of a candidate’s popularity or public support, particularly when shared without timestamps or location details.
Hannington Mutabazi, an editor with Vision Group, explains the verification challenges this environment creates.
“This election period is very sensitive because people are divided on whom they support. What makes it even more challenging is that we are in an era of artificial intelligence. Everything we use or get, we must double-check to make sure it is accurate, " Mutabazi says.
"Many people are going to create fake videos and fake audios to push an agenda, and it is journalists’ role to verify these before giving them a voice or risk misinforming the masses,” he added.
Mutabazi notes that AI-assisted editing tools are increasingly embedded in commonly used software, making it harder to immediately identify manipulated images or videos, especially when journalists are working under time pressure.
“As the fourth estate, journalism should break the cycle and provide accurate information,” he asserts.
Newsroom pressure can lead to rushed decisions, but journalists emphasise the importance of verifying material using reliable sources and on-the-ground confirmation.
Mutabazi also points out that many African countries lack comprehensive digital databases, limiting the usefulness of AI chatbots for local verification.
“It would be a disservice for journalists to verify news or election outcomes from ChatGPT or any other AI model because these chatbots use data that may not be accurate.”
AI-driven misinformation takes several forms. Deepfake videos can fabricate events that never occurred, while image-generation tools can insert or remove crowds from rally footage to exaggerate or downplay public support.
Synthetic audio can replicate the voices of political figures, and AI-generated text can be used to create fabricated statements or documents that appear official when shared as screenshots.
When such material is distributed on platforms like WhatsApp, TikTok, or Facebook without context, it can influence perceptions of political momentum and legitimacy.
Audience Exposure and Interpretation
Uganda’s digital environment plays a role in how this content is received. With millions of smartphone users and a predominantly young population, social media remains a primary source of political information.
Urban audiences may have access to multiple news outlets and live coverage, enabling cross-checking. Rural audiences, however, often rely on forwarded videos, voice notes, and images shared by trusted contacts.
In these settings, AI-generated or manipulated visuals—such as rally crowd images—can be persuasive, particularly when they align with existing political beliefs.
Government officials have acknowledged that election periods heighten the risks associated with AI-enabled misinformation. Policy discussions within the Ministry of ICT and National Guidance have highlighted the need to balance technological innovation with safeguards against misuse.
While Uganda has legal frameworks addressing online communication, including provisions within the Computer Misuse Act, observers continue to debate how these laws apply to emerging forms of AI-generated content.
Mutabazi describes existing laws as preventative in intent.
"Laws are meant to prevent wrongdoing, not to punish, like the Computer Misuse Act. Government needs to do more awareness activities so that young people know right from wrong," Mutabazi argues.
Verification Pressures in Newsrooms
Journalists covering the 2026 elections report receiving an increasing volume of videos and images claiming to show rallies, speeches, or crowd reactions. Verifying whether such content is current, authentic, or manipulated often requires contacting reporters on the ground, analysing metadata, or comparing footage across sources.
Rhonet Atwine, a journalist covering digital misinformation, highlights the risks posed by synthetic content during tense political moments.
“My biggest worry is that AI-made speeches can be used to fake leaders’ voices and words… If not checked early, they can mislead voters, cause fear and anger, and even push people into violence before the truth is known,” Atwine says.
Michael Ainomugisha notes that even experienced journalists face delays when verification becomes more complex.
“I mainly face challenges differentiating what is real and what is not… This delays the process of disseminating information and often leads to back-and-forth discussions with editors,” Ainomugisha says.
A senior political reporter at a Kampala-based newsroom described how synthetic content has reshaped daily reporting routines.
“We now receive videos and voice notes that look and sound authentic, but verifying them can take hours. By the time you confirm the truth, the misinformation has already spread,” the journalist said.
Observers agree that AI-driven misinformation has become a practical challenge rather than a theoretical one.
As Uganda approaches the 2026 elections, the effectiveness of verification systems, legal interpretation, newsroom practices, and public understanding will influence how political information is consumed and trusted.
Whether AI amplifies confusion or is managed through coordinated institutional responses will depend largely on decisions made before voters go to the polls.