Your Face, Your Voice, Someone Else’s Words: India’s Deepfake Problem Is Now Personal
Imagine waking up to find a video of yourself circulating online. It has your face, sounds exactly like your voice and shows you saying things you never said.
Your family believes it. Your colleagues begin asking questions. Strangers abuse you online. By the time you explain that the video is fake, it has already travelled across social media and countless WhatsApp groups.
This is no longer a scene from a science-fiction film. It is the reality of the deepfake era.
Artificial intelligence can now imitate a person’s face, voice and expressions with disturbing accuracy. What once required expensive equipment and technical expertise can increasingly be done using easily available digital tools. The result is a new kind of misinformation one that does not merely tell a lie but makes it appear as if a real person is telling it.
When public figures become targets
The danger has returned to the spotlight after allegedly manipulated videos linked Union Minister Nitin Gadkari to the controversy surrounding ethanol-blended fuel.
The matter reached the Bombay High Court, which directed social media platforms to remove the disputed content. When some posts reportedly remained accessible, the court again asked the platforms to comply and explain the action they had taken.
Around the same time, an AI-generated video impersonating Tamil Nadu Chief Minister C. Joseph Vijay began circulating online. The video appeared to promote financial assistance and displayed numbers for people to contact. Citizens were warned not to call those numbers or send money.
The two cases reveal different sides of the same danger.
A deepfake can be used to damage someone’s reputation, influence a political debate or make a fraudulent scheme appear genuine. When viewers see a familiar face making an announcement, they are more likely to trust it. That trust is precisely what the creators of such videos exploit.
If it can happen to a minister, it can happen to anyone
Cases involving politicians attract public attention, police action and legal scrutiny. But most victims do not have that kind of reach.
A fake video could target a schoolteacher, student, doctor, employee or small business owner. It could show someone making an offensive statement, endorsing a fraudulent product or admitting to something they never did.
Even if the video is eventually proved false, the damage may not disappear with it. Relationships can be affected, careers can suffer and reputations built over years can be questioned within minutes.
This is what makes deepfakes deeply personal. The technology does not only threaten famous people. Anyone whose photographs or videos are available online could potentially be impersonated.
A lie only needs a few seconds
Social media rewards content that creates an immediate emotional reaction. A video that makes people angry, frightened or shocked is more likely to be shared before anyone checks whether it is authentic.
Most people do not pause to locate the original source. They may not check whether the statement also appears on the official account of the person concerned. A familiar face and a confident voice often feel like enough proof.
By the time a clarification is issued, the original video may have reached millions of people. The denial rarely travels as widely as the accusation.
That is the uncomfortable truth of the internet: a lie can complete its journey before the truth has even found its shoes.
What happens during an election?
Deepfakes become especially dangerous during elections.
A fabricated video released shortly before voting could show a candidate insulting a community, admitting to corruption or making a controversial promise. Even if the video is exposed later, some voters may already have made their decision.
India has introduced rules requiring political parties and candidates to clearly label AI-generated or digitally altered campaign material. Social media platforms are also expected to act quickly when unlawful or misleading content is formally brought to their attention.
But even a few hours can be a very long time online.
A video can be downloaded, copied and uploaded from multiple accounts within minutes. Removing the original post does not automatically erase the versions circulating on other platforms and private messaging groups.
The problem, therefore, is not only how quickly one video is removed. It is whether its copies can be traced and whether the people responsible for creating and deliberately spreading it can be identified.
India has rules, but implementation will matter
India has strengthened its Information Technology Rules to address deepfakes and other AI-generated content. The framework requires AI-altered material to be clearly labelled and calls for measures that can help trace its origin.
Social media companies are also expected to prevent the circulation of unlawful content that impersonates people, misleads users or causes harm.
These protections are necessary, but rules alone will not solve the problem. Their success will depend on how seriously platforms respond to complaints and how quickly they identify manipulated content.
An ordinary user should not have to fight a lengthy legal battle to have a fake video removed. Platforms need simple complaint systems, faster human review and clear explanations of the action taken.
They should also reveal how many complaints involving deepfakes they receive, how many posts they remove and why some complaints are rejected.
Not every AI-generated video is harmful
Artificial intelligence is not automatically the villain.
AI-generated videos can be used responsibly in cinema, education, art, satire and entertainment. A clearly labelled parody is very different from a video designed to make people believe that a real person said or did something.
The problem begins when technology is used without consent to deceive, defame or defraud.
Clear labelling is therefore essential. If a video has been generated or significantly altered using AI, viewers should be told in a way that is easy to notice. A warning hidden in tiny letters or buried in a caption serves little purpose.
The label should also remain attached when the video is downloaded and uploaded somewhere else.
How can an ordinary person spot a deepfake?
Earlier deepfakes often contained visible mistakes: unnatural blinking, awkward lip movement, strange lighting or a robotic voice. Those clues still help, but the technology is improving rapidly. A sophisticated fake may not contain any obvious visual error.
The safest approach is to examine the source rather than trusting appearances.
If a video contains a sensational announcement, check whether it appears on the person’s official account or on an official government website. Look closely at the account that uploaded it. Newly created pages, unusual usernames and videos without a clear source should raise suspicion.
Be especially careful if the video promises jobs, prizes, financial assistance or government benefits and then asks you to call a number, click a link or send money.
Most importantly, do not forward something merely because it looks convincing. If you are uncertain, pause.
Social media platforms cannot remain silent spectators
Technology companies often say that users create and upload the material appearing on their platforms. That is true, but users do not decide which video will be recommended to millions of people. Algorithms do.
Platforms profit from engagement, and sensational content usually attracts more of it. If a false video is being actively recommended because it keeps people watching, the platform cannot completely distance itself from the consequences.
Companies controlling the distribution of information must share responsibility for preventing deliberate deception. They need stronger detection systems, visible warnings and tougher action against accounts that repeatedly circulate manipulated material.
When even the truth can be dismissed as fake
Perhaps the greatest danger of deepfakes is not that people will believe every fake video. It is that they may eventually stop believing genuine videos as well.
A public figure confronted with an authentic recording could simply claim that it was created by AI. Real evidence of wrongdoing could be dismissed as a deepfake. The same technology that manufactures false proof can also create doubt about genuine proof.
Democracy does not require everyone to hold the same opinion. It does, however, require some shared understanding of facts.
If society can no longer agree on who said what, what really happened or whether a video is genuine, meaningful public debate becomes almost impossible.
The question is no longer whether artificial intelligence can manufacture a convincing lie. It can.
The real question is whether our laws, social media platforms, institutions and everyday habits can protect the truth before the lie becomes more believable than reality.