![FourthEstate](https://www.fact-checkghana.com/wp-content/uploads/2024/08/fcg-logo.png"){kind=logo}
Claim: A TikTok audio allegedly features Dr Mahamudu Bawumia, the NPP presidential candidate, calling Ghanaians weak-minded and urging his supporters to convince the electorates with more promises.
Verdict: False! Our analysis revealed significant discrepancies between the viral audio and Dr Bawumia’s verified public speeches. The audio was generated using voice-cloning technology.
Background
Ahead of Ghana’s upcoming election, a TikTok audio allegedly of Mahamudu Bawumia began to make the rounds. In the viral audio, Dr Bawumia, the Vice President of Ghana and a presidential candidate under the New Patriotic Party (NPP), made disparaging remarks about Ghanaians, acknowledging past deceit and promising further manipulation in a bid to win the election. The audio, shared with his image, has sparked widespread reactions, with many questioning its authenticity.
“I believe it is very possible. Ghanaians are gullible and weak-minded. Most of them have even forgotten the lies we told them in 2016. So, guys, let’s stick to that script and keep promising them a lot of goodies. After we win, we’ll teach them a lesson,” the speaker partly said.
Apart from TikTok, the audio has also appeared on Facebook as seen here and here.
Verification
The Ghana Fact-checking Coalition compared the viral TikTok audio to five of Mr Bawumia’s past speeches on YouTube, revealing significant discrepancies. The audio features unnatural pitch variations and a polished accent that differs from his typical speaking style, suggesting manipulation.
The Hiya Deepfake Voice Detector tool also rated the audio as 80 percent likely to be a deepfake. Experts from the Ghana Fact-checking Coalition concluded that the audio likely stems from voice-cloning technology, such as Resemble AI, which can generate synthetic speech from text scripts.
Conclusion
The claim is false. Discrepancies between the viral audio and Mr Bawumia’s verified public speeches indicate it was generated using voice-cloning technology.
