A video of a well known politician saying something shocking. A voice note that sounds exactly like your boss asking for an urgent Mobile Money transfer. A screenshot of a "government notice" that never existed. Artificial intelligence has made these fakes cheap and easy to produce. Knowing how to spot deepfakes and AI generated fake news is now a basic life skill for anyone in Uganda who uses WhatsApp, Facebook, TikTok or X.
What is a deepfake?
A deepfake is a piece of media, usually video, audio or an image, that has been created or altered with AI to make it look as if someone said or did something they did not. The term comes from "deep learning", the type of AI used. Related forms of fake content include:
- Voice clones: AI that imitates a person's voice from a short recording.
- Face swaps: placing one person's face on another person's body in a video.
- Generated images: pictures of events, crowds or disasters that never happened.
- AI written articles: convincing fake news stories, sometimes on websites designed to look like real media houses.
- Edited documents: fake letters, circulars, receipts and certificates.
Why deepfakes matter in Uganda
Messages spread fast through family and community WhatsApp groups, and people tend to trust what friends forward. That makes fake content effective for:
- Scams: fraudsters impersonating relatives, managers or bank staff to request money.
- Political misinformation: especially around election periods.
- Reputation attacks: fake videos or images targeting businesses, public figures or ordinary people.
- Health rumours: false cures and alarming claims during outbreaks.
Sharing false information can also have legal and social consequences, so it pays to check before forwarding.
How to spot deepfake videos
AI video keeps improving, so no single sign is proof. Look for several clues together:
- Look at the face edges. Blurring or flickering around the hairline, ears and jaw can indicate a face swap.
- Watch the eyes and blinking. Unnatural blinking, a fixed stare or eyes that do not track properly are warning signs.
- Check lip sync. Do the lip movements match the words exactly, especially on sounds like "p", "b" and "m"?
- Notice the lighting. Shadows on the face that do not match the room, or skin that looks unusually smooth.
- Study the hands and teeth. AI often struggles with fingers, jewellery and the detail of teeth.
- Listen to the background. Sudden changes in background noise or an unnaturally clean sound can suggest editing.
How to spot fake voice notes
Voice cloning scams are particularly dangerous because many Ugandans use voice notes daily. Warning signs include:
- A familiar voice with a strange tone, flat emotion or odd pauses.
- An unusual request, especially for money, PINs, codes or secrecy.
- Pressure to act immediately: "Send it now, I will explain later."
- A new or unknown number claiming the person "lost their phone".
The best defence is simple: hang up and call the person back on the number you already have saved. Agree a family or office code word that must be used for any urgent money request.
How to spot AI generated images
- Zoom in on text in the image. AI often produces garbled signboards, number plates and labels.
- Check hands, ears and accessories for strange shapes.
- Look for repeated patterns in crowds, buildings or trees.
- Notice an overly polished, dreamlike look.
- Run a reverse image search using Google Lens or a similar tool. If the photo appears years earlier in a different context, it is being misused.
How to check fake news articles and screenshots
- Check the source. Is it from a known media house's official website or verified page? Look closely at the web address for misspellings.
- Search the headline. If something big happened, several reputable outlets will report it.
- Check the date. Old stories are often recirculated as if new.
- Look for official confirmation. For government notices, check the ministry or agency's official website or verified social media accounts.
- Be careful with screenshots. Screenshots of tweets, letters and chats are easy to fake. Look for the original post.
- Read beyond the headline. Fake stories often have sensational headlines with little detail, poor grammar or no named sources.
A simple pause before you share
Before forwarding anything that makes you angry, scared or excited, ask:
- Who created this, and how do I know?
- What evidence supports it?
- What do other trusted sources say?
- What happens if it is false and I share it?
If you cannot answer confidently, do not share. Strong emotions are exactly what fake news is designed to trigger.
Tools that can help
- Reverse image search: Google Lens, Bing Visual Search and TinEye.
- Fact checking organisations: several African fact checking groups publish checks on viral claims. Search for the claim with the words "fact check".
- Platform labels: some social networks label AI generated or altered media. Treat labels as helpful but not complete.
- AI detection tools: these exist but are not reliable on their own. Use them as one clue among many.
What businesses should do
Companies are targets too. A fake video of a "CEO announcement" or a cloned voice instructing the accounts team to pay a supplier can cost millions of shillings. Protect your organisation by:
- Requiring call back verification on a known number for any payment change or urgent transfer request.
- Using a professional email domain so customers can recognise genuine messages, and training staff to check sender addresses.
- Publishing official announcements on your own website so people have a trusted place to confirm news.
- Responding quickly and publicly if a fake involving your brand circulates.
Frequently asked questions
Can deepfakes be detected with certainty?
Not always. The best fakes can fool experts. That is why checking the source and context matters more than studying pixels.
What should I do if someone makes a deepfake of me?
Save evidence such as links, screenshots and dates, report the content to the platform, inform people close to you, and consider reporting to the police if it involves threats, fraud or harassment.
Is it illegal to share fake news in Uganda?
Uganda has laws covering computer misuse and related offences, and sharing false or harmful content can carry consequences. Laws change, so seek proper legal advice for specific situations. The safest habit is to verify before you share.
Are all AI generated images bad?
No. AI images are widely used for design, advertising and art. The problem is when they are presented as real to mislead people.
Conclusion
Deepfakes and AI generated fake news will only become more convincing, but a few habits go a long way: check the source, verify through a second channel, look for visual and audio clues, and pause before sharing. Teach these steps to family members and staff, especially anyone who handles money. A little scepticism protects your wallet, your reputation and your community.