By Benedict AbuBakarr Conteh
One of the AI-generated images published by Biggy Smalls on Facebook
In recent months, a Facebook user operating under the name Biggy Smalls has attracted widespread attention for sharing breathtaking photos purportedly taken in various locations across Sierra Leone and some parts of Africa, Asia, Europe and America.
These images include majestic mountains, futuristic buildings, crystal-clear rivers snaking through dense green hills, highways, bus stations in Freetown and the provinces, the Lungi Bridge, which is still under the planning stage, Victoria Park, villages, amongst others that appear untouched by modern life. Many of the images are captioned as “beautiful Sierra Leone”, “Good morning from Freetown”, “greetings from Sierra Leone”, “Freetown bus terminal”. “Lungi bridge” and “good evening from Mile 91”.
But here’s the problem: none of them are real.
To the average online user, the posts feel like a celebration of the country’s hidden beauty and cultural heritage. To anyone with local knowledge or OSINT training, it feels like something else entirely.
What at first appears to be an effort to promote tourism is, in reality, a campaign built almost entirely on falsehood using AI-generated visuals. The images are fabrications captured not with a camera, but with algorithms from Midjourney, DALL·E, Meta AI or other text-to-image generators.
DUBAWA’s investigation reveals that the Facebook account is creating or reposting AI-generated images, falsely claiming that they are authentic photos of real places in Sierra Leone. Biggy Smalls, knowingly or unknowingly, is constructing a fictional version of Sierra Leone’s landscapes, and this seemingly harmless deception could have long-term consequences.
And for the first time, our researcher can confirm that some of these images were created using Meta AI, and thus turning Facebook into both the platform and the tool of misinformation.
The Facebook user, Biggy Smalls, has been active on the social media platform since June 3, 2020, amassing over 230,000 followers with a steady stream of posts. Its identity is murky. A few personal photos. No tagged friends. No traceable history, only surreal landscapes tagged “Good morning from Sierra Leone.”
The online persona “Biggy Smalls” presents himself as a public figure based in Glasgow, United Kingdom, where he claimed to have grown up. His profile offers limited personal information aside from his studies at Glasgow Clyde College between 2020 and 2023, where he claimed to have graduated. He describes himself often using motivational statements like “No Condition Is Permanent” and “Keep Going, you’ll surely get there…”, and thus projecting an aspirational image.
Initially, the social links provided on the profile, including two Instagram accounts and a Threads account, lead to dead ends, thus raising concerns about their authenticity. However, further investigation using the keywords “Biggy Smalls” on Instagram and Threads uncovered two active accounts tied to the persona. These accounts are posting content that corroborates the activity on the original Facebook profile, indicating a deliberate cross-platform presence. They are @sheku197 on Threads and @sheku197 on Instagram.
Screenshot of the profiles that match Biggy Smalls Facebook profile.
DUBAWA investigation also revealed that all these accounts on Facebook, Instagram, Threads, and Audiomack are tied to the common Sierra Leonean name “Sheku” or “Sheku Koroma.” This cross-linking of identities raises further suspicion about the nature of the persona and its intentions.
Audiomack profile of Biggy Smalls that was shared via the Biggy Smalls Facebook profile.
Biggy Smalls shares an Audiomack account containing several music tracks. Our analysis confirmed that all these tracks were created using the AI music platform Suno, rather than being originally produced compositions, complicating his claim of being a music creator.
The first clue: Too perfect to be real
It started with a post on April 19, 2025. “Good morning from Sierra Leone. 🇸🇱”
First AI-generated photo posted by Biggy Smalls
Above is an image showing a vibrant outdoor gathering of people dressed in colourful traditional attire, holding and waving multiple Sierra Leonean flags against a bright, partly cloudy sky. This digitally created scene conveys a sense of national pride and festivity but does not represent a real event or factual situation.
The composition includes men and women in a variety of patterned garments and headwear, standing closely together in a celebratory manner, with flags prominently displayed in the foreground and background. While visually representative of a public celebration, the image was produced using Meta AI and should be interpreted as a simulated depiction rather than a photographic record of an actual occurrence.
Our researcher ran the image through Illuminarty AI detection tool and returned an AI probability score of 97.1%. Manual examination of the image shows Meta AI watermark at the extreme right of the photo.
Screenshot showing Illuminarty AI probability rating of 97.1%
The pattern of fabrication
From that day, we began examining more posts from the same user.
Infographics showing AI-generated photos posted by Biggy Smalls since April 19 to July 16, 2025.
Between April 19 to July 16, 2025, Biggy Smalls published over 90 images allegedly showing scenes from Sierra Leone.
A few of these photos include:
An overhead of a bridge, labeled as “Sierra Leone ”.
A busy community of people and a flag of Sierra Leone supposedly in Mile 91 Sierra Leone.
A police station claimed to be from Yenga, Sierra Leone .
A photo of a bus station showing “Kambia Sierra Leone ”.
A village showing small houses and a flag of Sierra Leone in one of the houses, claimed to be “Bo Town Sierra Leone .”
A border crossing point showing military personnels, a riverline and a sign board showing Yenga Sierra Leone.
An airport tagged “Beautiful Sierra Leone Lungi International Airport”.
A photo of a well decorated hospital tagged “Sierra Leone Government Hospital Kenema”
A standard image of a bank with people in the streets walking tagged, “Bo Town Sierra Leone ”.
A bridge showing “Lungi Bridge Sweet Mama Salone ”
Amongst these photos, a few have been fact-checked by DUBAWA as seen here, here, here, here and here.
After DUBAWA debunked these AI-generated claims, Biggy Smalls didn’t stop posting images that have the possibility of disinforming his over 230,000 followers on Facebook; in fact, he increased the trend by sharing AI-generated scenery images of other countries.
Amongst them are photos tagged: Tanzania, Jamaica High School , Nigeria , Senegal, Kenya , Nigeria Sokoto State, Ethiopia and India .
To hide the Meta AI label on the photos and portray them as realistic, he placed an emoji on most of the images, especially wherever the AI labels appeared. Some were cropped whilst others still have the Meta AI label on them.
Screenshot of the Analysis by Forensically magnifier tool on Biggy Smalls photo
For instance, when this image was analyzed using the Forensically Magnifier tool, it detected that it has an AI label similar to that of Meta AI.
This AI-generated image depicts a lion standing majestically on a rocky ledge overlooking a brightly lit cityscape at night, with the Sierra Leonean flag positioned prominently nearby. The composition blends natural symbolism and urban imagery, with the lion representing strength and resilience while the illuminated city, complete with light trails and modern high-rises, suggests progress and vitality. The presence of the Sierra Leonean flag underscores a theme of national pride, though the scene is a digitally created artistic visual and does not depict a real location or event.
One of the AI-generated photos posted by Biggy Smalls tagged Good afternoon from Sierra Leone
Biggy Smalls posted another image on July 4, 2025 at approximately 4:18 PM, with a caption thus: “Good afternoon from Sierra Leone The Lion Mountain ”. Since posted, it has gained 7.9K, Likes, 211 Comments and 37 Shares on Facebook.
Screenshot of the post made by Biggy Smalls on July 4, 2025.
We used the Fotoforensics Metadata tool to extract the metadata and analyze the image.
Screenshot of the analysis made with Fotoforensics on the image.
Every one of those identified images was confirmed by AI-detection tools as synthetic, some generated by Midjourney with remnants of distortions on structures and shadows, Bing Image Creator or DALL·E and Meta AI.
Most posts were made between 7 AM and 10 AM GMT, while others were posted around 4 PM and 6 PM GMT and almost all used vague generic topics like: “Sierra Leone”, “Good morning Sierra Leone”, “Good morning Freetown”, “Mile 91”, “Good evening Freetown” etc.
Meta AI-generated images often exhibit distinctive visual traits, including blended or dreamlike skies, over-saturated green landscapes, soft edges with fog overlays, repeated or identical trees and structures, and unrealistic lighting with unnatural depth-of-field effects.
The real-world Impact
While sharing beautiful images of a country may seem harmless, the consequences of synthetic geography are significant. Such imagery fuels misinformation, leading locals and diaspora communities to believe these fictional places exist, often resulting in disappointment or a sense of betrayal when reality doesn’t match expectations. It also opens the door to tourism-related fraud, where individuals or businesses could exploit fake imagery to attract funds, investments, or visitors to non-existent locations. Over time, this practice erodes public trust, as people grow skepticalAll efforts to reach out to Biggy Smalls on Facebook in relation to our investigations have proven unsuccessful, as our researcher is unable to send a message via his profile.iggy Smalls’ posts, many users expressed doubt and curiosity, asking questions such as, ‘Where exactly is this in Salone? I’ve never seen it,’ and ‘This can’t be in Bo — is it from Europe?’ Despite these concerns, most of these questions remain unanswered. The page owner neither clarifies the origin of the images nor labels them as AI-generated, leaving viewers with assumptions and confusion. This silence suggests either a deliberate attempt to mislead or, at the very least, a troubling negligence in maintaining digital integrity. To gather and analyze these public reactions, our researcher employed a comment extractor tool called Export Comments.
Some extracted comments from two posts made by Biggy Smalls tagged good evening Freetown and Freetown Highway.
Ophaniel Gooding, one of the coordinators at the Sierra Leone Association of Journalists SLAJ iVerify, highlighted the significant threat AI-generated images pose to real-world situations, especially in a country like Sierra Leone.
“The danger is huge…AI-generated images can be deliberately used to spread tribalism, regionalism, or even incite conflict. They can also distort public opinion on important issues like elections, the judiciary, or the executive, which affects decision-making in our democratic process.”
Ophaniel advised: “Always verify. Whether you’re literate or illiterate, before sharing or acting on any information you see online.”
Also speaking with DUBAWA, Murtala Mohamed Kamara, Founder and CEO of Jamboree Consult, a digital marketing agency in Sierra Leone, notes the increasing difficulty in verifying information, even for professionals, due to AI tools that can clone voices and create realistic videos. To avoid falling for false content, he advises verifying the source, cross-checking with multiple platforms, reporting misleading posts, and staying informed through digital literacy and critical thinking.
“The deliberate use of AI-generated images to misrepresent geographical locations is a deceptive and unethical tactic that misleads and misinforms the public. This practice can negatively portray a place, making it seem unappealing, particularly to those unfamiliar with it,” Kamara said.
Furthermore, Abass Sesay, a journalist and development advocate in Sierra Leone, highlights that this deceptive behaviour is not exclusive to social media influencers but is also employed by some journalists.
This case highlights a subtler yet equally disturbing form of disinformation, one that doesn’t revolve around falsehoods about people or politics but instead fabricates a sense of place. Unlike traditional fake news, which often exploits fear, anger, or hate, this type of content romanticizes reality, painting idyllic yet false images of a nation. However, the outcome remains the same: audiences form false beliefs based on synthetic creations.
When asked if he knows that he’s misinforming the public with his posts and the dangers it poses to other social media users, Biggy Smalls responded: “Are you a Facebook police officer? What’s wrong with you? Stop sending me messages.”
Screenshots showing Biggy Smalls reply to our researcher when asked about his posts and the dangers it poses to other social media users.
While Sierra Leone currently has no specific legislation regulating the use or attribution of AI-generated images, existing provisions within the Independent Media Commission Regulations 2022 may be applied in certain contexts. Notably, Section 6 prohibits the broadcast of untrue information.
However, these laws were not specifically designed for synthetic media such as AI-generated visuals, and their application remains limited by technical and institutional enforcement challenges.
Meta, the parent company of Facebook and Instagram, has implemented several mechanisms to address the misuse of AI-generated content, including its Community Standards, partnerships with third-party fact-checkers, and the use of automated detection tools. These tools aim to identify and reduce the visibility of manipulated or misleading content. However, enforcement remains inconsistent, particularly in underrepresented regions like Sierra Leone, largely relying on user reports or partner intervention.
Although Meta’s policy requires labelling of manipulated media and synthetic images that could deceive the public, it does not always automatically detect geographically misattributed AI-generated visuals unless they are flagged or go viral. Users can report misleading posts through built-in tools, but platform-level action depends heavily on volume and verification.
0 Comments