A video shows three well-known Nepali political leaders quietly meeting together, sparking rumours of a secret alliance. A relative calls in a panicked voice claiming to be in trouble and asking for money urgently. A video of a respected businessperson promoting a "guaranteed" investment scheme circulates on Facebook. In each case, the person you think you are seeing or hearing may not have said or done anything at all - it may be a deepfake.
Deepfakes - AI-generated video, audio, or images designed to convincingly imitate a real person - have moved well beyond internet novelty. During Nepal's 2026 election cycle, fabricated videos of political figures spread widely enough that fact-checkers had to formally debunk them, and researchers monitoring the campaign described the country as facing a genuine "digital battleground" of manipulated content. The same underlying technology that produced those political fakes is now being adapted, across South Asia, into financial and personal scams.
This guide explains how these scams typically work, the warning signs to watch for, and exactly what to do if you encounter one.
If a video, voice call, or image is asking you to send money, share an OTP, or act urgently - especially if it involves a "famous" or "trusted" person you cannot independently verify is really speaking to you - treat it as a likely scam until proven otherwise. Real emergencies and real investment opportunities do not require you to skip verification.
Why This Is Becoming a Bigger Problem in 2026
Deepfake-related fraud is not a fringe issue anymore. Global identity-fraud researchers tracked a sharp rise in AI-driven identity fraud through 2026, with document and video-based deepfake attacks growing far faster than traditional fraud methods, largely because generative AI tools have made convincing fakes possible with just one photo or a short prompt rather than hours of specialised editing work.
Nepal's own experience so far has centred mainly on political disinformation - fabricated videos of leaders and manipulated photos of candidates circulated during the election period. But the same tools used to fake a political meeting can just as easily fake a celebrity endorsing an investment scheme or a family member in distress, which is exactly the pattern already being reported elsewhere in the region.
How a Typical Deepfake Scam Unfolds
Common Deepfake Scam Formats to Know
1. Fake celebrity or CEO investment endorsements
Scammers generate a realistic video of a well-known businessperson, banker, or public figure "recommending" a stock, crypto, or trading platform with guaranteed returns. Victims are directed to a fake trading app or website where their deposits simply disappear.
In a widely reported case in India, a chartered accountant lost approximately Rs 23 lakh (roughly 2.3 million Indian rupees) after being convinced by deepfake videos of well-known business figures promoting a stock investment scheme on social media. While this specific case occurred in India rather than Nepal, it illustrates a scam pattern that uses the exact same AI tools now circulating across South Asia, including Nepal.
2. Cloned "family emergency" voice calls
A caller's voice is cloned from a short public audio or video clip - sometimes from a Facebook video or a recorded call - and used to convincingly imitate a relative in distress, urgently requesting money for a hospital bill, bail, or an emergency transfer.
3. Fake job offers and video interviews
Scammers use AI-generated "HR representatives" or manipulated video calls to run convincing fake interviews, eventually asking candidates to pay a "processing" or "training" fee for a job that does not exist.
4. Political and social disinformation adjacent to scams
Nepal has already seen confirmed cases of fabricated videos of political figures and AI-generated clips falsely linked to real unrest, verified as synthetic by independent fact-checkers using AI-detection tools. While these are not always direct financial scams, the same techniques and easy availability of AI tools that produced them are the same tools scammers use for fraud.
5. Fabricated compromising images for extortion
AI image tools can be used to generate fake compromising or explicit images of a real person, which are then used to demand money under threat of "sharing" the fabricated content. This is a serious form of extortion and, importantly, a criminal offence.
Red Flags That Signal a Possible Deepfake Scam
| Warning Sign | Why It Matters |
|---|---|
| Urgency - "act in the next few minutes," "limited slots" | Scammers rely on panic to prevent you from verifying independently |
| Request to move to a private channel quickly (WhatsApp, Telegram) | Removes the content from a platform that might flag or remove it |
| Unnatural blinking, lighting, or lip-sync mismatch in a video | Common technical giveaway in lower-quality deepfakes |
| A voice that sounds slightly flat, robotic, or has odd pacing | Voice cloning often struggles with natural emotional tone |
| A "guaranteed" or "risk-free" investment return | No legitimate investment can honestly guarantee returns |
| Payment requested via personal wallet, crypto, or untraceable transfer | Legitimate institutions rarely request payment this way |
| Refusal or inability to video call back on a separate, verified channel | A real person or institution can always be reached another way |
How to Verify Suspicious Content Before You Act
- Call the person back directly on their known number - never the number that contacted you
- Ask a question only the real person would know the answer to
- Search the claimed investment platform's name plus "scam" or "review" before sending any money
- Check whether the same video or claim appears on the public figure's own verified official page
- Slow down - a five-minute delay to verify has never caused a real emergency to get worse
- Look closely at hands, ears, teeth, and background details in videos, where AI artifacts are most common
- Be suspicious of any request for an OTP, PIN, or password, regardless of who appears to be asking
Agree on a simple "family safe word" with close relatives that would never appear in a public video or social media post. If someone claiming to be family cannot say it during an urgent call, treat the call as a scam.
What to Do If You Have Been Targeted or Scammed
Do not send any additional money "to recover" a previous loss - this is a common follow-up scam targeting people who already lost money once.
Save the video or image file, screenshots of the conversation, phone numbers, wallet or bank account details used, and exact timestamps before anything is deleted.
If a transfer was made, report it right away - faster reporting meaningfully increases the chance of freezing or reversing a transaction.
Complaints can be filed online through their portal, by email, or in person at their office in Bhotahity, Kathmandu. District police offices can also accept and forward complaints.
Facebook, WhatsApp, TikTok, and YouTube all have reporting options for impersonation and manipulated media - use them to request removal.
Scammers frequently target multiple people connected to the same victim or public figure - a quick warning can stop the next attempt.
The Legal Angle in Nepal
Deepfake-based fraud, impersonation, and extortion in Nepal generally fall under the Electronic Transactions Act, 2063 (2008), which addresses offences including unauthorised access, dissemination of illegal or obscene content, identity theft, and email or online fraud, alongside relevant provisions of the National Penal Code covering fraud, forgery, and privacy violations. The Cyber Bureau of Nepal Police is the designated authority for investigating these cases, working alongside local district police offices. If you are dealing with extortion involving fabricated intimate images, this is treated as a serious offence, and you should report it rather than pay - paying rarely stops the threat and often invites further demands.
This section is general information, not legal advice. If you are facing an active extortion attempt or a significant financial loss, consult a qualified lawyer alongside filing your police complaint.
Frequently Asked Questions
Look for unnatural blinking patterns, mismatched lip movement and audio, inconsistent lighting or shadows on the face versus the background, and blurred or warped details around the ears, hair edges, and hands. Newer deepfakes are harder to spot visually, so verifying through an independent channel is more reliable than visual inspection alone.
Yes. Modern voice-cloning tools can produce a convincing imitation from just a short public sample, such as a Facebook video or a recorded phone call, which is why urgent voice calls asking for money should always be verified through a separate, known channel.
Using a deepfake to defraud, impersonate, harass, or extort someone falls under offences addressed by the Electronic Transactions Act and relevant penal code provisions in Nepal. Enforcement depends on reporting the case to the Cyber Bureau with clear evidence.
Recovery is not guaranteed, but reporting immediately to your bank and the Cyber Bureau gives you the best possible chance, particularly if the transfer can still be traced or frozen. Never pay an additional "recovery fee" to anyone who contacts you promising to get your money back - this is a common secondary scam.
No. While celebrity-impersonation investment scams often target people with money to invest, family-emergency voice scams and extortion attempts using fabricated images can target anyone, regardless of income or public profile.
Report to the Cyber Bureau, Nepal Police, either through their online complaint portal, by email, or in person at their office in Bhotahity, Kathmandu. You can also file a complaint at your nearest district police office, which will forward it to the Cyber Bureau.
Conclusion
Deepfakes are no longer a distant, futuristic threat - they are already circulating in Nepal's political space, and the same tools are being adapted into financial and personal scams across the region. The most reliable defence is not spotting every technical flaw in a fake video, since the technology keeps improving. It is building a habit of slowing down at the moment of urgency, verifying through a separate and trusted channel, and refusing to send money, OTPs, or personal information based on a video or voice call alone. That single habit defeats the vast majority of deepfake scams, regardless of how convincing the fake looks.
Before you trust the next AI-transformed photo you see online, know what actually happens to a photo once you upload it.
Read our guide: Is It Safe to Use AI Image Generators With Your Photo in Nepal? ->
Discussion