A few weeks ago, when an audio said to be of journalist Solomon Sserwanja appeared online, a fact checking organization for which his African Institute of Investigative Journalism (AIIJ) is a partner, quickly dismissed the recording as fake. A few weeks later, the so-called leaked video of Western Uganda youth MP aspirant Elizabeth Kakwanzi Katanywa appeared online. It, too, was dismissed by members of the Kakwanzi camp.
The Sserwanja ‘audio’ and the so-called Kakwanzi ‘leaked video’ touched on sens!t!ve matters of the heart.
In the Solomon Sserwanja case, the audio carried a voice of a man (that sounded like Sserwanja) love chatting up a woman (thought to be a side dish).
Both media were dismissed as having been generated by artificial intelligence.
The African Center for Media Center (ACME) has offered some insights on how to detect an artificial intelligence generated audio or video.
First, some videos or audios may be deep fakes that might be difficult for some to detect.
“In the case of a voice deepfake, AI technology is used to replicate someone’s unique speech patterns, accent, tone and even emotional inflections, making it sound as though they said something they never actually did,” explained ACME, a Ugandan organization bridging the space between the media industry and journalism training institutions.
“Voice deepfakes are particularly concerning because they are easier and faster to create than video deepfakes, and are often more difficult for listeners to scrutinise, especially in casual listening environments. They can be convincing even with short audio samples, and their impact is amplified on platforms like WhatsApp or X (formerly Twitter), where audio quality may already be degraded or compressed. This makes it harder for ordinary users to determine whether the voice is genuine or artificially generated.”
ACME guides that checking checking the origin of the video or audio can help one to judge if it is original or deep fake. This involves checking the person or people who posted the audio or video and if there are any accompanying screenshots or contextual text.
For audios, the African Centre for Media Excellence further guides, one can examine the recording using tools like DeepFake-o-meter, AI Voice Detecter and Resemble Detect “to view the audio waveform.”
“Do you see unnatural gaps or repeated patterns? In higher-end programmes, you can check spectrograms (visual sound maps) for missing frequencies or signs of artificial voice generation. Listen for unusual cadence, lack of emotion or speech that sounds oddly clipped,” it added.
One also seek human clues.
ACME advises that one plays the audio to people familiar with the speaker’s real voice to see if they recognise it. “Are there inconsistencies in the language, accent or phrasing? Does the emotional tone match the context? Does the speaker sound too calm or too detached?”
The organization also advises reporters and other content creators to reach out to people said to be in the audios or so-called videos to see if they deny or confirm being in the media in question.
Another tip from the African Centre for Media Excellence is looking for “supporting or contradictory evidence” related to the video or audio. This may include “messages, videos or photos that back up the events described” or “eyewitnesses or other records that support or challenge the story.”
It should be remembered that in the most recent case of youth MP aspirant Kakwanzi there were even claims that she had leaked her own video to gain political popularity. (See related reporting Here, There and Over There).
For comments on this report, story tips or sponsored content, send us a Whatsapp message on +256 705 690 819 or E-mail us on pearltimesug@gmail.com).
Please Follow our Whatsapp Channel for More Stories Here
