How a Nigerian fact-checking platform used isFake.ai to expose two deepfake prison videos. Part 1

DUBAWA, one of Africa’s leading fact-checking organizations, analyzed two viral videos of Nnamdi Kanu and confirmed they were AI-generated — using IsFake.ai as one of the key verification tools.

Dec 4, 2025
Preview
0:00
0:00

How fact-checkers and AI caught a political deepfake red-handed. Part 1

December 2025. Two videos hit Facebook, claiming to show activist Nnamdi Kanu in a Nigerian prison. They spread fast with thousands of shares, heated comments, and real outrage. But some commenters were concerned. The team at DUBAWA decided to investigate. And they didn’t just trust their eyes. They used our AI video detector to ask the hard question: is this video real or fake?

Turns out, both clips were digital fiction. Our detector gave Video #1 a 72% AI-generated score and Video #2 63%. The investigation revealed classic deepfake giveaways: weird crowd glitches, a spoon serving nothing onto a full plate, and even Kanu’s signature facial scar was missing. Oops.

The deepfake breakdown

Here’s what DUBAWA found when they dug into the footage:

  • distorted background figures showing the typical ‘AI crowd glitch’,
  • a spoon serving no food, while the plate is already filled,
  • misaligned hand and utensil movements,
  • missing facial scar on Kanu (a known physical identifier),
  • unnatural motion artifacts.

DUBAWA submitted both videos to isFake AI video checker.Results left no doubts:

  • Video #1: 72% AI-generated visual elements,
  • Video #2: 63% AI-generated visual elements.
isFake AI flagged both clips as digitally altered, with strong indicators of face-replacement and synthetic frame reconstruction. This matched the organization’s manual review and supported their final verdict.

To validate the analysis, DUBAWA also used Attestiv and Hive Moderation tools. isFake.ai served as a key cross-validation point for confirming manipulation.

Why this case matters

This wasn’t just a silly fake. It was a politically charged deepfake aimed at influencing public opinion during a tense moment. It shows how AI-generated misinformation can look convincing enough to spark real reactions — until someone checks properly.

Source: Dubawa fact-check report & Original Facebook reel

FAQ

Is this prison video of Nnamdi Kanu real?

No. Combined fact-checking and our AI video detector confirm it's predominantly synthetic (72% AI-generated). It’s a deepfake video designed to mimic real footage.

Why would someone create a fake prison video?

To manipulate narratives. In tense political climates, a convincing AI-generated video can inflame supporters, discredit authorities, or create false evidence. All to influence public perception and real-world actions.

How do fact-checkers and AI detectors work together?

Fact-checkers investigate context, sources, and claims. AI tools like our video detector analyze the digital file for technical signs of synthesis. Together, they provide a complete debunking.

What’s the danger of political deepfakes like this?

They undermine trust in institutions, can incite violence, and make it impossible to discern truth. When AI-generated misinformation looks real, it becomes a weapon against social stability.

What should I do if I see a suspicious political video?

Do not share it immediately. Look for reputable fact-checking reports (like Dubawa’s), check the upload source, and consider running it through an AI video detection tool. Breaking the chain of AI-generated fake news starts with skeptical sharing.

FEED YOUR VIDEO TO THE DETECTOR

From $5.99/month

Sign In
Premium