DEEPFAKES : The Thin Line Between Seeing and Believeing

“I’ll believe it when I see it… is about to be rendered obsolete thanks to recent technological innovations enabled by artificial intelligence (AI).” The difference between real and artificial is suddenly becoming a thin line. For those that are not tech savvy this places them at a very bad place in this undeniable industrial revolution.?

Even if you thought Africa is always the last to cope up with new systems , the AI advent and its potential benefits and risks is a nightmare even in the rural areas yet to receive 5G internet. For myself the worry is , is my grandmother aware that there is deepfakes and it could mislead an entire village??

Deepfakes are synthetic media? created using AI algorithms.Anyone with access to the Internet – with just a few minutes of effort and minimal skill – can create powerful videos.These tools are not only available to governments and private companies – but more recently to the public.

Months ago around the time of U.S president Donald Trump inauguration, a deep fake video trended.The trending deep fake video claimed that former Kenyan Deputy President Rigathi Gachagua had attended Donald Trump's inauguration. However, this was flagged across the internet as? false confirming that Gachagua did not attend the inauguration, and the video was a fabrication.

AFP Fact Check conducted reverse image searches on screen grabs from the TikTok clip to find the original footage. They found the altered video was based on a broadcast of The Ingraham Angle show aired by Fox News on January 17, 2025 ,host Laura Ingraham .

So while leaving the office for my field assignments , the first thing the driver asked me and my camera man was “ is it true Gachagua attended Trump’s inauguration ?” and then boom! It dawned on us that finally “ the deep fake age is here with us!”

So we told him “ dere (slang for driver)no, he didn’t attend , the video was made up by the latest technological advancements called Artificial Intelligence.”

While it is essential to verify such information before believing or sharing it, a single share can cause damage across the masses on the internet. It doesn’t require too many shares but just one that amasses too many other reshares to make the damage done and dusted. While the Gachagua one is just one among the many that have trended and the others that await to rain on our mobile screens.

The most recent case is Kenya's Foreign Affairs Principal Secretary Korir Sing'Oei, who? recently issued an apology after mistakenly posting a deep fake video that was deceptively labeled as a CNN report.

On February 20, 2025, Korir Sing’oei posted a video on his official X account that appeared to be a CNN broadcast featuring journalist Fareed Zakaria praising Kenya’s diplomatic efforts in mediating the ongoing Sudan conflict. Sing’oei captioned it: “For the fake analysts who doubt the good faith of Kenya’s peace diplomacy, here is a cogent assessment by CNN’s Foreign Policy expert, Fareed Zakaria .” The video was intended to support Kenya’s credibility amid criticism over its hosting of Sudan’s Rapid Support Forces (RSF) for peace talks, a move that had drawn scrutiny from Sudan’s government and others questioning Kenya’s neutrality.

However, the video was quickly identified by Kenyans on X and media outlets as an AI-generated deepfake—not a genuine CNN segment. No such broadcast existed on Zakaria’s official platforms, and the production showed telltale signs of artificial manipulation.

Mr. Sing’oei even so pledged to undertake training in artificial intelligence to avoid similar incidents in the future. However my question stands ; does an apology mend the damage ? and how then can we sensitize and create awareness around AI’s potentiality of spreading disinformation an social untruths ?

?Here are ways to Combat Deepfakes :

1. Build Awareness and Digital Literacy

Educate people about deepfakes—what they are, how they’re made, and their risks. Teach critical thinking to question media authenticity.

Informed people are less likely to fall for fakes, as seen when Kenyans on X quickly debunked Sing’oei’s video.

2. Establish Verification Protocols

Set up clear methods to confirm media or requests, especially for sensitive actions (e.g., financial transfers, official statements).

?It adds a human layer of skepticism, breaking the trust deepfakes exploit.

3. Leverage AI Detection Tools

Use AI to fight AI—deploy software that analyzes media for manipulation signs (e.g., pixel irregularities, synthetic audio patterns).

Tech can scale detection beyond human limits, catching subtle flaws we miss.

4. Implement Labeling and Watermarking

Require AI-generated content to carry visible markers (e.g., watermarks, metadata tags) to signal it’s synthetic.

Transparency reduces deception risk, letting users judge content knowingly.

5. Strengthen Platform Policies

Push social media and video platforms to detect and remove deepfakes or flag them prominently.

Platforms are key spreaders—controlling them limits reach.

6. Encourage Public Vigilance and Reporting

Mobilize communities to call out fakes and share verification habits.

Collective action outpaces individual efforts, overwhelming deepfake spreaders.

7. Limit Personal Data Exposure

Reduce the raw material (photos, videos, voice clips) deepfakes need by tightening privacy.

Thus starves deepfake creators of inputs

8. Legislate and Regulate

Push for laws to penalize malicious deepfake creation and mandate detection tools.

?Legal teeth back up tech and education, hitting intent where it starts.

9. Foster AI Research and Innovation

Invest in better detection and anti-deepfake tech through public-private collaboration.

It keeps defenses ahead of AI’s evolving threats.

Article | By Isaac Oketch