The Eight-Second Test, Will India’s New IT Rules Contain Deepfakes in an Age of Viral Deception and Shrinking Attention?
The comprehensive amendment to the Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021, announced on February 10 by the Ministry of Electronics and Information Technology, is a welcome step in the era of AI-generated content. The rules, which come into effect on February 20, clearly define “synthetically generated audio, visual or audio-visual information” using computer resources and specify mechanisms to contain the generation and dissemination of false and malevolent information, often referred to as deepfakes.
The accompanying analysis by V. Sridhar, a professor at IIIT-Bangalore, offers a critical examination of these rules and raises troubling questions about their effectiveness. It acknowledges the importance of the new regulations but argues that they face fundamental challenges: the shrinking attention spans of users, the incentive structures of social media platforms, the sophistication of deepfake technology, and the difficulty of enforcement against malevolent actors who will not comply with labelling mandates.
The rules prescribe the due diligence to be observed by intermediaries in containing the production and dissemination of deepfakes. They specify that intermediaries should inform users to prominently label such synthetically generated information for easy identification. While empowering intermediaries to monitor user content, the rules also endow them with associated responsibilities and liabilities for hosting or propagating deepfakes. However, the power bestowed on intermediaries to flag and take down synthetically generated information suo motu is likely to hit some guardrails.
The Creative Potential and the Dark Side
In this age of generative AI, the creation of synthetically generated information has become dramatically easier with the availability of hundreds of tools. Nvidia-backed AI voice startup ElevenLabs, which provides voice and video cloning apps, recently hit a $13 billion valuation. These apps are being used by creators to combine their work with AI-driven synthetic media to generate performances for the next generation.
These AI tools have dramatically reduced the entry barriers for new and upcoming artists, writers, and performers, allowing them to create niche markets for themselves amidst the giants. These and other similar creative uses of synthetically generated information are broadly permitted under the new rules. The labelling mandate is meant to curb information created in bad faith.
But malevolent actors will not label their content. They are not interested in complying with rules; they are interested in spreading deception. A comprehensive audit mechanism is required to book such offenders. This is a law enforcement challenge that the rules alone cannot solve.
The Virality Problem: Why Deepfakes Spread
What bothers governments all over the world is the virality of deepfakes that are violent, abusive, malevolent, and obscene. Researchers have found that deepfakes often are more viral due to their innovative creation and visually compelling nature. A study indicates that deepfake-based cyber attacks or frauds happen once every five minutes.
Social media platforms are the primary sources for encountering and disseminating deepfakes. Hence, the additional responsibility placed on significant social media platforms for checking the authenticity of labels on synthetically generated content is a welcome measure indeed. However, whether they will act on deepfakes that are viral at the cost of losing eyeballs is a moot question.
This is the platform dilemma. Social media platforms are driven by engagement. Their algorithms reward content that captures attention, generates reactions, and keeps users scrolling. Deepfakes, by their very nature, are often visually compelling, emotionally charged, and highly shareable. They are designed to go viral. A platform that aggressively removes viral deepfakes risks alienating users and losing market share to competitors that are less scrupulous.
The Attention Challenge: Eight Seconds and Counting
While prominent labelling should help identify deepfakes, will it be effective given the short attention span of today’s users, especially the youth? It is estimated that the attention span on social media of Gen Z is eight seconds, compared to 12 seconds of millennials. This is one of the primary reasons for the popularity of short video formats and reels.
In an experiment conducted with students at IIIT, the author found that even tech-savvy students were not able to fully identify deepfakes that were created using AI tools, as they so closely mimic originals. If experts struggle, what hope is there for the average user? A small label in the corner of a video is unlikely to be noticed, let alone heeded, when content is consumed in quick, superficial bursts.
This is not a failure of individual users; it is a feature of the medium. Social media is designed for speed, not depth. It rewards the immediate over the considered, the emotional over the rational. A labelling mandate that relies on user attention to be effective is fighting against the very architecture of the platforms it seeks to regulate.
The Educational Imperative: Building Deepfake Literacy
The experiment at IIIT also offered a ray of hope. Researchers found that educational intervention—providing tips on detecting deepfakes—resulted in marked improvement in detection rates. This suggests that while labelling alone may not be sufficient, it can be part of a broader strategy that includes awareness building and digital literacy.
The author argues that, apart from the well-founded IT rules, the government should also create an awareness campaign for alerting netizens to deepfakes. Regulations introduce constraints in the use of technology and hence have a negative effect on innovation. But awareness campaigns have no such downside. They empower users without burdening platforms or stifling creativity. They are a low-cost, high-impact complement to the labelling mandate.
Such campaigns should target not only tech-savvy youth but also semi-literate and older populations, who may be even more vulnerable to deception. The goal is not to make everyone an expert but to cultivate a healthy scepticism, to encourage users to pause and question before sharing.
The Limits of Regulation: Why Enforcement Matters
The analysis also highlights the limits of regulation. Malevolent actors will not label their content. They will not comply with the rules. A comprehensive audit mechanism is required to identify and prosecute such offenders. This is a law enforcement challenge that requires resources, expertise, and international cooperation.
Deepfakes often originate outside India, using servers in other jurisdictions. Tracing them, identifying the perpetrators, and bringing them to justice is a complex and time-consuming process. The government must invest in the capacity to do this effectively. Without such enforcement, the labelling mandate will only catch those who choose to comply, leaving the worst offenders untouched.
Conclusion: Beyond Labelling
India’s amended IT rules are a significant step forward in the effort to contain deepfakes. They recognise the right of users to know the origin of the media they consume and place responsibility on platforms to facilitate this transparency. But the analysis makes clear that labelling alone is not enough.
The effectiveness of the rules depends on factors beyond their scope: the shrinking attention spans of users, the incentive structures of social media platforms, the sophistication of deepfake technology, and the capacity of law enforcement to pursue malevolent actors. Addressing these challenges requires a multi-pronged strategy that includes awareness campaigns, digital literacy programmes, platform accountability, and robust enforcement.
The eight-second attention span is not going to lengthen. Deepfakes are not going to become easier to detect. The only variable that can change is our collective response. The government, platforms, educators, and users must work together to build a digital ecosystem that is resilient to deception. The labelling mandate is a beginning, not an end.
Q&A Section
Q1: What are the key provisions of India’s amended IT rules regarding synthetic content, and what is their intended purpose?
A1: The amended IT rules, which come into effect on February 20, 2026, clearly define “synthetically generated audio, visual or audio-visual information” using computer resources and specify mechanisms to contain the generation and dissemination of false and malevolent information (deepfakes). The rules prescribe due diligence for intermediaries, require them to inform users to prominently label synthetic content for easy identification, and empower intermediaries to monitor user content while assigning them responsibilities and liabilities for hosting or propagating deepfakes. The intended purpose is to curb the spread of harmful synthetic content while permitting legitimate creative uses. However, the analysis notes that the power bestowed on intermediaries to flag and take down content suo motu may hit guardrails, and malevolent actors will not comply with labelling mandates, requiring a comprehensive audit mechanism for enforcement.
Q2: What is the “platform dilemma” regarding viral deepfakes, and why does it complicate enforcement?
A2: The “platform dilemma” refers to the inherent tension between platforms’ economic incentives and their regulatory responsibilities. Social media platforms are driven by engagement; their algorithms reward content that captures attention, generates reactions, and keeps users scrolling. Deepfakes, by their very nature, are often visually compelling, emotionally charged, and highly shareable. They are designed to go viral. A platform that aggressively removes viral deepfakes risks alienating users and losing market share to competitors that are less scrupulous. Yet the new rules place additional responsibility on significant social media platforms for checking the authenticity of labels on synthetic content. The dilemma is that the same platforms expected to police synthetic content are economically dependent on the engagement that such content generates. The analysis does not resolve this dilemma but highlights the need for a fundamental rethinking of platform incentives, not just additional regulation.
Q3: Why is the eight-second attention span of Gen Z a fundamental challenge for the effectiveness of labelling?
A3: The eight-second attention span means that content is consumed in quick, superficial bursts, with little time for reflection or verification. A small label in the corner of a video is unlikely to be noticed, let alone heeded. Users are focused on being entertained, informed, or outraged, not on verifying authenticity. The label becomes an obstacle to the user’s goal, an intrusion that disrupts the flow. Even if noticed, it may be ignored. This is not a failure of individual users; it is a feature of the medium. Social media is designed for speed, not depth. A labelling mandate that relies on user attention to be effective is fighting against the very architecture of the platforms it seeks to regulate. The analysis argues that this challenge cannot be overcome by regulation alone; it requires a broader strategy that includes awareness building and digital literacy.
Q4: What role can educational interventions play in addressing the deepfake challenge, according to the IIIT experiment?
A4: The IIIT experiment found that educational intervention—providing tips on detecting deepfakes—resulted in marked improvement in detection rates among students. Even tech-savvy students initially struggled to identify deepfakes created using AI tools, as they so closely mimic originals. However, after receiving targeted training, their ability to detect deception improved significantly. This suggests that awareness campaigns and digital literacy programmes can be highly effective complements to regulation. The analysis recommends that the government create awareness campaigns targeting not only tech-savvy youth but also semi-literate and older populations, who may be even more vulnerable to deception. Unlike regulations, which can constrain innovation, awareness campaigns have no downside. They empower users without burdening platforms or stifling creativity.
Q5: Why is a comprehensive audit mechanism necessary, and what challenges does it face?
A5: A comprehensive audit mechanism is necessary because malevolent actors will not label their content. They are not interested in complying with rules; they are interested in spreading deception. Identifying and prosecuting these offenders requires a system that can trace deepfakes to their source, gather evidence, and coordinate with law enforcement. This is a complex challenge for several reasons. Deepfakes often originate outside India, using servers in other jurisdictions, requiring international cooperation. Tracing them requires technical expertise and resources that may be lacking. And the legal process is slow, while deepfakes spread in hours. The analysis argues that the government must invest in the capacity to pursue malevolent actors effectively. This includes training law enforcement, building technical capabilities, and establishing international partnerships. Without such an audit mechanism, the labelling mandate will only catch those who choose to comply, leaving the worst offenders untouched.
