The advent of deepfake technology, which uses artificial intelligence to create manipulated videos and images that appear authentic, has given rise to a major societal concern in India. With the proliferation of social media, deepfakes can spread rapidly and cause serious harm before the victim even realizes what is happening.
The recent incident where a deepfake video of popular actress Rashmika Mandanna was circulated online has brought this issue under sharp focus. Even though Mandanna was quick to point out that it was a fake, the potential for damage was clear. Experts say that deepfakes can be used for revenge porn, identity theft, financial fraud, character assassination and other nefarious purposes. Women are particularly vulnerable to deepfakes being used to harass them.
With general elections coming up, there are worries that deepfakes could be used to spread political misinformation and manipulate voters. Fake speeches or media interviews of politicians are some ways deepfakes could wreak havoc. The danger of deepfakes promoting hate speech and inciting violence is also real.
Cybersecurity researchers have warned that deepfake technology is evolving rapidly while awareness about it is still low. India lacks robust laws and regulations to deal with this novel threat. The Information Technology (Intermediaries Guidelines and Digital Media Ethics Code) Rules, 2021 contain some provisions for social media platforms to curb misinformation, but experts say more needs to be done.
The government recently issued an advisory directing social media platforms like Facebook, Instagram and X to take down deepfake videos within 24 hours of receiving a complaint. Failure to comply could invite punitive action under IT laws. It has also advised victims to file police complaints.
While this is a welcome move, lawmakers need to think long-term instead of acting in a reactive, ad-hoc manner. A comprehensive legal framework to counter the spread of deepfakes is the need of the hour.
Regulating Deepfakes: The Legal Landscape
Deepfakes fundamentally challenge existing laws which rely on content being demonstrably false to consider it illegal. Since deepfakes are designed to look genuine, they occupy a gray area with respect to veracity.
Simply banning deepfakes through new laws may be counterproductive since the technology also has many constructive applications and stifling innovation in AI is unwise. However, unregulated deepfakes pose too much of a threat to shrug off. Lawmakers have to walk a fine line between protecting citizens’ rights and nourishing technological progress.
India’s IT Act Already Has Provisions Against Impersonation
Section 66D of the Information Technology Act contains penalties for identity theft and cheating by impersonation using a computer resource. This covers types of deepfakes like face-swapping. But while this provision has existed since 2008, deepfakes themselves are a relatively new phenomenon that have exposed its inadequacies.
The anonymity afforded by the internet makes tracking down deepfake creators difficult. And most victims do not even realize a deepfake exists until the damage is done. So while the law offers retrospective penalization, more work needs to be done for deterrence and prevention.
Social Media Platforms Need Clearer Guidelines
The current government advisory asks platforms like Facebook and Instagram to remove reported deepfake videos within 24 hours or face action under IT rules. However, the rules themselves do not define clear-cut procedures for online intermediaries to deal with deepfakes proactively.
The self-regulatory Code of Ethics under the IT rules does make general statements about curbing misinformation but lacks specific guidelines. Critics argue that an issue as serious as deepfakes warrants proper standard operating protocols for platforms baked right into the IT rules. Relying on platforms to act responsibly under broad principles is not enough.
One area where clearer guidance is needed is the turnaround time for evaluating takedown requests. The 24-hour deadline set arbitrarily by the government runs the risk of letting some deepfakes slip through if platforms feel pressured to act hastily without proper verification. A reasonable timeframe for review balanced with the urgency to limit viral spread should be prescribed through an amendment.
Other jurisdictions like Australia have passed laws holding platforms liable if they do not expeditiously take down harmful deepfake content after being notified. India could consider similar accountability measures that encourage proactive monitoring by platforms instead of a reactive attitude. However, cooperation and consultation with the tech industry is advisable rather than a draconian crackdown.
Anti-Deepfake Strategies Used Globally
Many countries are grappling with the complex policy issues surrounding deepfakes. Analysing some of their regulatory approaches can help India draft its own blueprint.
– The US has introduced the Deepfake Report Act that requires companies developing deepfake detection tools to evaluate their systems’ strengths and weaknesses annually and share their findings with the government. This promotes transparency and accountability.
– China has simply outright banned deepfake apps like Zao from its internet. But critics call China’s approach an excuse to increase online censorship since even harmless deepfakes get blocked.
– South Korea has introduced a law requiring deepfake creators to disclose their creations prominently. This reduces the tendency to pass off deepfakes as real. However, enforcing disclosure norms on anonymous creators online poses challenges.
– Singapore has recently criminalized distribution of intimate deepfake media without consent. But this is a limited solution focusing only on one specific harm caused by deepfakes.
– The European Union is considering far-reaching laws like obliging platforms to get creators’ consent before publishing synthetic media featuring them. However, this prior restraint approach is prone to over-regulation.
A balanced regulatory approach for India should likely involve elements of disclosure norms, mandatory take-downs, cooperation with platforms as well as AI innovations like forensic watermarking. But the debate on the appropriate policy mix must include stakeholders from government, industry, civil society, academia and technology. Knee-jerk reactions without public discourse should be avoided.
Why Relying on AI is Not Enough
Deepfake creation and deepfake detection are both AI techniques racing against each other. Some believe allowing this adversarial AI vs AI contest to play out is the optimal solution. If detection AI becomes good enough, the deepfake problem will supposedly resolve itself without heavy-handed regulations.
However, relying solely on next-gen AI is myopic. While counter technologies are important, they cannot be the first line of defence. Once a deepfake spreads online, the damage is often irreversible. And AI is still years away from being able to detect all deepfakes flawlessly.
Prevention is better than cure. The onus should be on lawmakers to erect stringent barriers against malicious uses of deepfakes through legislation. Simultaneously, platforms must be obligated to actively seek and remove deepfakes instead of acting only after complaints. AI can augment this framework by making enforcement more robust but cannot replace regulations and human monitoring.
The Road Ahead for India
Deepfakes represent a major inflection point for truth and trust in the digital age. Even as technology marches ahead, human values and ethics should not be undermined. With its strong IT sector and tech-savvy youth demographic, India must not lag behind in tackling the policy quandaries created by deepfakes.
The government has taken the first step by acknowledging the issue and advising remedial actions. But a long-term vision backed by legislation is indispensable. Constructive public debate can shape a progressive regulatory framework that reconciles innovation and prevention.
Specific priority areas include defining clear obligations for platforms, nurturing detection technologies, upgrading cyber-related laws and enhancing cybersecurity resources. Holistic international collaboration will also be key since deepfakes transcend national boundaries.
Deepfakes are a boiling frog scenario — a slowly emerging threat that will likely catch society unawares unless acted upon preemptively. India today has a vital window of opportunity to get ahead of the curve. With visionary leadership and public participation, we can preempt this ticking time bomb before it exacts a devastating toll. The time for prolonged deliberation is over. Concrete action is the pressing need of the hour to secure India’s digital future.