By Abdulahmeed Abdulrozaq Bilal | Fact Frontier.
Click here to read more about the reporter.
A political firestorm has erupted online following the circulation of a highly inflammatory audio clip, artificially generated to mimic the voice of President Bola Tinubu, with presidential aide Bayo Onanuga leading the charge for the immediate arrest of controversial social media influencer Martins Otse, popularly known as VeryDarkMan (VDM).
The controversy centers on a fabricated voice note that has been universally condemned as a product of Artificial Intelligence (AI), yet its dangerously seditious contents have triggered widespread debate, forcing the presidency to threaten legal action against anyone connected to its distribution.
What Was in the Fake Audio?
Despite being debunked as a deepfake, the audio clip succeeded in capturing national attention due to the gravity of the statements falsely attributed to the Commander-in-Chief. Investigations into the manipulated clip revealed a disturbing narrative designed to portray the government as a malicious entity working against the Nigerian people.
In the fake recording, the AI-generated voice of President Tinubu allegedly made a series of confessions regarding national insecurity. Most alarmingly, the clip claimed that the administration was deliberately allowing the spate of kidnappings and killings in the South-East to persist as a political weapon. The voice stated that the government had decided to let the region “taste its own medicine” for its political choices in the last general election.
Furthermore, the fabricated audio took aim at the opposition. The voice claimed that President Tinubu had personally told Labour Party’s presidential candidate, Peter Obi, to “step down or be dealt with,” insinuating a direct threat from the presidency to the opposition leader. It also suggested that the President was completely indifferent to the economic hardship facing citizens, allegedly dismissing the masses' suffering with callous remarks.
In another segment designed to incite public outrage over the country’s rising debt profile, the AI-generated voice falsely claimed that loans recently secured from the World Bank were not intended for infrastructure or palliatives but were being secretly diverted to fund upcoming elections and silence judicial officers.
The Connection to VDM and the Arrest Warrant
The crisis deepened when Bayo Onanuga, Special Adviser to the President on Information and Strategy, directly linked VDM to the dissemination of this fake audio. Onanuga stated that VDM acted as the “conveyor and disseminator” of the falsehood, insisting that “VDM needs to face the weight of the law.” This statement instantly ignited the trending call for his arrest, with security agencies reportedly monitoring the situation.
The fabricated audio is believed to have originated from a manipulated segment of one of VDM’s own original videos. Unidentified creators isolated a clip of VDM speaking, then superimposed an AI voice-clone of Tinubu over it, making it appear as if the President’s voice was leaking from VDM’s device. While VDM may not be the original creator of the deepfake, the presidency maintains that sharing such seditious content—which alleges government-sponsored terrorism and electoral fraud—constitutes a clear violation of cybercrime laws.
Reacting to the arrest threats, VDM remained characteristically defiant. In a response video, he dared the government to act swiftly. “I will even book a flight to come earlier than I planned. I want to see what you people can do,” he said, arguing that the government was prioritizing the persecution of online critics over solving genuine security crises.
As the presidency doubles down on identifying the origin of the AI manipulation, the incident has sparked a national conversation about the regulation of artificial intelligence and the limits of free speech, leaving the digital space divided between those demanding VDM’s prosecution and those calling for an independent verification of the increasingly blurred lines between fact and deepfake.
