|
Getting your Trinity Audio player ready...
|
Introduction
The emergence of artificial intelligence (AI)-powered tools has changed Ghana’s information ecosystem. Bolstered by the availability of voice cloning, deepfake media, and synthetic video manipulation tools, the once-visible lines between manipulated and trusted content have been fundamentally blurred.
As manipulated media spreads across traditional and social media, distinguishing authentic public discourse from manufactured deception has become a challenge for journalists, businesses, governments, democratic institutions, and the public.
Assessing this worrying trend, DUBAWA critically examines how tech-assisted misinformation and disinformation are produced and distributed in Ghana, based on an empirical review of 43 verified fact-check investigations published between April 2025 and Sept. 2026. The reviewed reports revealed that misinformation and disinformation in the West African country extend far beyond casual political satire or isolated hoaxes. The study disclosed that synthetic media is actively deployed to manufacture credible-looking endorsements, invent foreign visits by global figures, fabricate political praise, simulate media interviews, and push fraudulent investment schemes to unsuspecting citizens.
Scope and methodology used
DUBAWA reviewed 43 fact-check reports, including 14 published between April and July 2025 and 29 published between Jan. and Sept. 2026, focusing on five key dimensions in its analytical framework.
(a) Format and technique – here the study focused on fully computer-generated media, face-swapping, voice cloning, audio manipulation, and repurposing authentic video with altered context;
(b) Subject matter – the subject matter of the posts involves domestic politics, international diplomacy, celebrity culture, commercial endorsements, and crisis reporting;
(c) distribution mechanics – here the study singled out the platforms that have been used to spread such viral content, including Facebook, TikTok, YouTube, and X (formerly Twitter);
(d) target audience and incentives – here the study analysed the targeted audience and motive through engagement farming, financial fraud, reputation building, and political influence; and
(e) ecosystem impact – here the study addressed the impact on individuals: financial loss, public confusion, reputational harm to public figures, and erosion of trust in legitimate journalism.
Distribution of the investigated cases reported by DUBAWA
The sampled cases reported by DUBAWA showed a continuous presence of synthetic media on social media across the observation period, peaking mid-year.
In April 2025, three cases were recorded, focusing primarily on fabricated celebrity encounters, such as actors Kwaku Manu and Dr Likee with French international Mbappé, as well as synthetic royal or political praise. See this, this, and this for DUBAWA’s report.
Also, in May 2025, DUBAWA reported four cases that emerged, characterised by false international endorsements, including U.S. President Donald Trump praising Ghanaian President John Mahama and U.S. billionaire Elon Musk purportedly visiting Ghana, alongside other fraudulent cryptocurrency promotions. See this, this, this and this for DUBAWA’s report.
Then, in June 2025, DUBAWA reported five cases marked by sophisticated broadcast media impersonation, including manipulated interviews of journalist Umaru Sanda and Ghanaian politician and actor John Dumelo, fabricated geopolitical developments like a fake coup attempt in Cameroon, and other synthetic content. See this, this, this, this and this for DUBAWA’s report.
In July 2025, two cases were analysed, defined by targeted financial scams, including a manipulated video of Ghanaian President John Mahama promoting a malicious investment link and a fake gaming app attributed to musician Sarkodie.
However, early 2026 saw a spike in synthetic content on Ghana’s social media. From one reported case in Jan. to two cases in both Feb. and Mar., to a record eleven cases in June. See this, this, this, this, this, this and this for some of the cases DUBAWA reported in June 2026.
Major Trends Identified by the study
The study identified six major trends from the 43 fact-check reports published by DUBAWA. These are fabricated endorsements by influential personalities, manufactured international political validation, impersonation of credible news formats and journalists, geopolitical crisis simulation and pan-African narratives, domestic politics and agenda setting, and financial fraud and malicious scheme promotions.
An infographic of the six major trends identified by DUBAWA’s study. This was generated using ChatGPT.
Fabricated endorsements by influential personalities
A primary strategy identified across the sample is the unauthorised deployment of globally recognised figures to validate local political, economic, or cultural narratives. Documented cases include fabricated videos of Donald Trump praising the Asantehene, as well as separate manipulated clips showing President Trump commending President John Mahama and Finance Minister Cassiel Ato Forson over cedi performance. Other cases feature edited clips claiming Elon Musk announced an impending visit to Ghana to meet MP Sam George and establish a Starlink regional headquarters, alongside manipulated media depicting American artist Beyoncé endorsing Burkinabè leader Ibrahim Traoré.
By leveraging international prominence, content creators create cognitive shortcuts that make extraordinary claims appear verified to everyday social media users.
Manufactured international political validation
Another trend identified in the study is manufacturing external political validation for regional leaders. Fabricated videos show foreign heads of state, billionaires or global icons praising local politicians or domestic governance. This trend allows creators to tap into national pride and partisan sympathies, granting unearned credibility to political campaigns and processes. See this, this, this and this for such synthetic content.
Impersonation of credible news formats and journalists
Perhaps one of the most dangerous trends is the exact replication of established media formats, particularly logos and fact cards. In June 2025, DUBAWA debunked a computer-manipulated interview featuring Citi FM and TV journalist Umaru Sanda Amadu and politician John Dumelo promoting a high-yield financial scheme. Similarly, in July, a fake Joy News segment impersonated rapper Sarkodie to promote a cash-prize gaming application.
Financial fraud and malicious scheme promotion
DUBAWA’s investigation revealed that a significant proportion of synthetic media in Ghana serves financial ends. The study identified false announcements attributed to public figures that frequently redirect viewers toward external links, betting portals, cryptocurrency schemes, or phishing sites designed to capture personal data and banking credentials.
Geopolitical crisis simulation and pan-African narratives
Several analysed cases involved complex geopolitical narratives, particularly surrounding Burkina Faso’s leadership and regional security. Examples include synthetic reports claiming Burkina Faso had built an aircraft manufacturing plant, fake letters from Ibrahim Traoré to Mike Tyson, and a fabricated video declaring a military coup against Cameroon’s President Paul Biya. These posts exploit heightened regional sensitivities, anti-colonial sentiment, and pan-African aspirations to maximise viral reach.
Domestic politics and agenda setting
DUBAWA’s study showed that some people generate content to create narratives that are not true in the country. This is done through the use of purported letterheads of either of the two dominating political parties in Ghana, the governing National Democratic Congress (NDC) and the opposition New Patriotic Party (NPP). See this, this and this for such narratives.
Target audience based on the 2025 and 2026 dataset
The analysis of the 2025 and 2026 dataset reveals that synthetic media in Ghana is carefully tailored to exploit the psychological triggers, emotional biases, and daily information habits of specific demographic groups. This conclusion was arrived at through an inductive content analysis of the 43 DUBAWA Ghana fact-check reports spanning the study period. The study revealed that creators design content around the pre-existing beliefs, fears, and interests of distinct target audiences in the country, including highly partisan voters and grassroots members, financially vulnerable Ghanaians, Pan-Africanists and youth activists, religiously motivated Ghanaians, and casual internet users.
An infographic of the target audience analysis identified by DUBAWA’s study. This was generated using ChatGPT
Platform Distribution and Media Manipulation Mechanics
The investigations revealed that synthetic deception relies on distinct social media platforms for different stages of content propagation. This was derived by conducting a platform-mapping and cross-channel traffic analysis across the 43 DUBAWA Ghana fact-check investigations from 2025 and 2026. The study uncovered that Facebook serves as the primary channel for high-reach page sharing and viral image circulation, as seen in the Kwaku Manu composite image and presidential investment scams. TikTok acts as the incubator for short-form video clips, face-swapping, and audio overlays, exemplified by the fake Elon Musk visit and the Burkina Faso manufacturing claim, which gathered over 142,000 likes at the time those stories were published. Also, YouTube facilitates longer-form, narrated sensationalism and crisis reports, such as the fake Cameroon coup video, which amassed over 530,000 views at the time of the story’s publication. Interestingly, X, formerly Twitter, accelerates rapid commentary, meme culture, and immediate public debate around celebrity encounters and media impersonations.
The study noted that creators deploy a spectrum of tools in the generation of these synthetic media, from simple image compositing and face replacement to advanced voice synthesis and the malicious re-contextualisation of authentic, historical video footage.
Ecosystem impact of synthetic media in Ghana
The systemic impact of synthetic misinformation on Ghana’s digital ecosystem operates across four key structural dimensions: information fatigue and “liar’s dividend,” institutional distrust and brand hijacking, economic vulnerability and financial scams, and aggravated social and political polarisation.
Information fatigue and the “liar’s dividend”
The widespread circulation of deepfakes induces widespread scepticism among the public. This creates a “liar’s dividend” where public figures and institutions can dismiss genuine, unflattering recordings or evidence as “AI-generated” without offering proof. This was inferred from cases in which public figures or political actors disputed the authenticity of genuine events by claiming that the media coverage was “AI-generated,” thereby making it difficult to distinguish genuine reports from fabricated content and allowing truthful reporting to be dismissed as synthetic. This local report documents how partisan actors copied and misused “FAKE NEWS” verification stamps issued by Ghanaian fact-checking platforms such as DUBAWA to cover up genuine news, thereby weaponising verification tools to generate further confusion.
Institutional distrust and brand hijacking
Manipulated videos impersonating major media outlets (e.g., Joy News, Citi TV) or public institutions damage institutional credibility. Traditional verification seals, official press releases, and “FAKE” stamps are routinely weaponised or duplicated by bad actors, diminishing their authority. This impact was inductively derived from the repeated occurrence of media impersonation in the dataset (e.g., deepfakes mimicking Joy News, Citi TV, and official government releases). Research has shown how exposure to deepfakes reduces general trust in news organisations, regardless of whether citizens believe the specific fake.
Economic vulnerability and financial scams
Generative voice clones and fabricated endorsements lower barriers for digital fraud. Fraudulent investment apps, fake bank branches, and impersonated public figures directly cause financial losses, targeting economically vulnerable citizens. This was extracted by synthesising the subset of fact-check reports involving financial fraud (such as PrimeAura, Ibrahim Mahama crypto schemes, and fake banking infrastructure). Researchers have highlighted how generative AI dramatically reduces the cost and technical barrier for producing highly persuasive financial and political fraud at scale. Specific empirical investigations document the direct financial claims, fraudulent links, and identity theft involving prominent leaders (e.g., Sammy Awuku, President Mahama) used to lure citizens into high-yield scam platforms.
Aggravated social and political polarisation
Synthetic content reinforces existing partisan confirmation bias and cultural anxieties. Fabricated claims surrounding political betrayals, judicial bias, or controversial social topics worsen political tribalism and lower the quality of civic engagement ahead of elections. Mapping the cases reviewed, DUBAWA identified synthetic media involving high-emotion societal topics (e.g., the Ningo-Prampram LGBTQ office image, partisan sabotage claims, and regional military coups). Researchers have revealed that digital misinformation rarely changes well-formed political beliefs. Instead, it reinforces strong confirmation bias and deepens existing societal divisions.
Limitations of the study
The limitations of the study stem primarily from the nature and scope of the underlying dataset. Owing to the limited number of stories, the study identified the following limitations: selection and publisher bias, lack of quantitative denominator, inconsistent reporting metrics, uncertain intent and attribution, and absence of direct audience perception data.
Selection and publisher bias
The research relies exclusively on reports investigated and published by DUBAWA Ghana. It does not capture the full, unmonitored volume of synthetic misinformation circulating across private messaging platforms (such as WhatsApp or Telegram groups) or unverified social media channels.
Lack of quantitative denominator
Without data on the total volume of all information circulating during the 2025–2026 period, it is impossible to calculate the exact percentage or national prevalence of digital misinformation that involves artificial intelligence versus traditional manipulation.
Inconsistent reporting metrics
Engagement data (views, likes, shares) across the fact-check reports were recorded at varying times after publication, limiting the ability to draw direct quantitative comparisons between different cases or platforms.
Uncertain intent and attribution
While the study identifies recurring narratives and mechanisms, the available fact-checks rarely provide definitive proof regarding the true identities, financial backing, or precise motivations of the content creators.
Absence of direct audience perception data
The fact-checks establish the deceptive claims and reach metrics, but they do not measure how many citizens actually believed the content, changed their political opinions, or suffered direct financial loss as a result.
Despite its methodological constraints, the study serves as a valuable framework for understanding the mechanics, distribution channels, and societal impacts of synthetic media in West Africa.
The study recommendations
One of the key highlights of the study is that the generation of synthetic media in Ghana’s media is growing by the day, and the time is ripe for some mitigating steps to be taken to remedy the situation. To mitigate the systemic risks posed by synthetic misinformation, DUBAWA recommends a multi-stakeholder strategy, involving fact-checking institutions, the media, platform owners, businesses and the government.
First, government institutions and public agencies must establish official, centralised digital verification hubs to rapidly confirm or deny statements attributed to the Presidency, state ministers, and security agencies. This is important because many of the reports reported targeted either the presidency or key government officials.
Second, social media platforms and tech companies need to enhance automated detection algorithms for accounts engaging in coordinated voice-cloning, face-swapping, and fraudulent investment promotion within West Africa. They must also provide clear, localised reporting pathways for users to flag manipulated media that impersonate journalists or news organisations.
Third, fact-checkers and media houses should standardise technical classification terminologies to distinguish clearly between fully synthetic media, manipulated audio, face-replacement, and simple contextual misuse. Building shared open-source databases of debunked audio tracks, source videos, and repeat-offender accounts will further streamline verification workflows across newsrooms.
Fourth, media literacy organisations and civil society ought to conduct public education campaigns emphasising that realistic speech, professional studio lighting, and recognisable faces no longer guarantee the authenticity of a video clip. Audiences must be trained to verify unexpected news or investment offers through official domain names and established media houses before sharing.
Conclusion
DUBAWA’s study of published fact-checks from April 2026 to Sept. 2026 highlights that tech-generated synthetic media is an active, evolving challenge within Ghana’s digital ecosystem. By fusing cutting-edge manipulation tools with familiar cultural icons, political figures, and news formats, malicious actors can generate widespread engagement and deploy sophisticated financial and political deception.
Addressing this threat requires a balanced approach. Strengthening technological verification, enforcing platform transparency, and deepening public digital literacy will safeguard Ghana’s democratic discourse and public trust without curtailing legitimate digital innovation.
