Last updated: 2 October 2026

AI-generated images, cloned voices and deepfake videos can now appear convincing enough to mislead ordinary viewers. A strange-looking face or unnatural voice may raise suspicion, but visual imperfections alone are no longer a reliable test.

The safest approach is to investigate the source, context and purpose of the content before believing or sharing it.

What Is a Deepfake?

A deepfake is synthetic or manipulated media created using artificial intelligence. It may depict a real person appearing to say or do something that never happened.

Synthetic content can include:

• AI-generated photographs of events that never occurred;

• manipulated videos using another person’s face;

• cloned voices imitating relatives, executives or public figures;

• fabricated celebrity endorsements;

• altered political speeches;

• non-consensual intimate images; and

• video calls designed to impersonate a trusted person.

Not every AI-generated image is harmful. Synthetic media can be used for entertainment, education and creative work. The danger arises when it is presented as authentic to deceive, defame, impersonate or obtain money.

Why Visual Clues Are Not Enough

Older deepfakes often contained obvious errors: distorted hands, unnatural blinking, mismatched lighting or poorly synchronised speech. These signs may still appear, but AI systems are improving quickly.

A video should not be declared fake merely because it looks unusual. Equally, realistic lighting and smooth speech do not prove that it is genuine.

Treat visual clues as reasons to investigate—not as final proof.

Seven Ways to Check Suspicious AI-Generated Content

1. Find the original source

Ask where the image, video or recording first appeared. A forwarded clip without an original post, date or identifiable creator should be treated cautiously.

Look for the content on the official website or verified account of the person or organisation concerned.

2. Search for independent confirmation

If the media appears to show an important announcement, public incident or controversial statement, check whether credible organisations have reported it.

A major event appearing only in one anonymous post is a warning sign.

3. Conduct a reverse-image search

A reverse-image search may reveal that a photograph is old, taken in another country or removed from its original context.

CERT-In recommends reverse-image searching as one method for examining a suspicious image or video thumbnail. However, a missing result does not establish authenticity.

4. Examine the complete frame

Look for inconsistent shadows, changing jewellery, distorted text, unnatural reflections, disappearing objects or irregular movement around the face and hair.

These details can indicate manipulation, but compression and poor connectivity can create similar effects.

5. Listen separately to the audio

Cloned speech may contain unusual pacing, missing breaths, repeated intonation or abrupt changes in background sound. Compare it with longer recordings from an official source.

If the voice is asking for money or confidential information, verify the request through another channel.

6. Look for labels and provenance information

Check whether the platform identifies the post as AI-generated, digitally enhanced or synthetic content.

India amended its IT Rules in February 2026 to strengthen requirements relating to synthetically generated information. The framework includes clearer labelling and traceable metadata for permissible AI-generated content.

Labels can help, but their absence should not be treated as proof that media is real.

7. Verify urgent requests directly

If a voice note or video call appears to come from a relative, employer or senior executive requesting money, end the interaction and contact that person through a number you already trust.

Consider agreeing on a private family or workplace verification phrase. Do not reveal that phrase publicly or through an unverified call.

Common Ways Deepfakes Are Misused

Financial impersonation

A cloned voice or video may imitate a family member, employer or public figure. The message creates urgency and requests money, credentials or confidential information.

Fake investment endorsements

Manipulated videos may show a celebrity, business leader or news presenter apparently promoting an investment platform. I4C has warned about sponsored advertisements using deepfake images and videos to direct people into fraudulent investment groups.

Political misinformation

Synthetic content can fabricate statements, events or endorsements. During elections, check official Election Commission communications and the original account of the person shown before sharing political media.

Non-consensual intimate imagery

AI can be used to create sexualised or intimate images without the person’s consent. Victims should preserve the relevant links and evidence without repeatedly downloading or circulating the material.

What to Do Before Sharing a Viral Image or Video

Pause and ask:

• Who first published it?

• Is the complete version available?

• Does the caption match the original date and location?

• Has the person or organisation responded?

• Is there independent reporting?

• Does the post request money or provoke immediate action?

• Would sharing it cause harm if it later proved false?

Do not repost suspicious content merely to ask whether it is genuine. Reposting can expand its reach even when accompanied by a warning.

Share a link to a fact-check or official clarification instead.

How to Report a Deepfake in India

Report the content to the platform on which it appears. Preserve:

• the complete URL;

• the account name and profile link;

• screenshots showing the post and date;

• the original message or email;

• associated phone numbers;

• transaction details if money was transferred; and

• the time and manner in which the content was received.

If the deepfake is connected to impersonation, threats, financial fraud, identity misuse or non-consensual imagery, submit a complaint at the National Cyber Crime Reporting Portal.

If money has been lost through a deepfake-assisted scam, call 1930 immediately and notify the relevant bank or payment provider. Read India News Grid’s guide on how to report cyber fraud in India.

For an immediate threat to personal safety, contact the local police rather than relying only on an online report.

The Bottom Line

There is no single visual trick that can identify every AI-generated image or deepfake video.

Check the original source, search for independent confirmation, examine the complete media and verify urgent requests through a trusted channel. Most importantly, do not allow a convincing face or familiar voice to replace independent verification.

If a suspicious video call includes threats of arrest or demands for money, also read India News Grid’s guide to digital arrest scams.

Official sources: CERT-In deepfake advisory, MeitY information on synthetic-content rules, National Cyber Crime Reporting Portal