By -Prisha Gakhar
Abstract
Ever since artificial intelligence has drastically changed political communication, the advent of deepfakes has also posed unprecedented threats to democratic processes. Deepfakes differ from traditional political misinformation by exploiting the reliability of audio-visual evidence itself, distorting the truth of the content and easier to weaponise during elections. In this article, I discuss the rise of AI-generated electoral rhetoric, focusing on India’s precarious position in this regard and analyze the country’s present regulatory framework, arguing that the dependence on patchy legislation and executive advisories is no match for the difficulties faced by AI-driven misinformation.
Introduction
The simple premise that democracy rests on is that citizens can believe what they see and hear. Political debate, conflicting views, and even misinformation are not new phenomena. What is new, however, is the presence of technology that may recreate reality itself. This article looks at whether India’s existing legal framework is ready to tackle the peculiar threats posed by AI-generated deepfakes during elections. It argues that its existing architecture is still fragmented and reactive, leading to substantial gaps in protecting electoral integrity. The article follows a doctrinal and comparative approach. It examines current deepfake use in election settings, looks at India’s current legal and regulatory framework, and offers a brief comparison of India’s approach with emerging international trends, especially the EU’s AI Act, to ascertain if existing law can handle AI-driven electoral manipulation and suggests a way forward.
The Rise of Deepfakes and the New Information Ecosystem
In July 2024, a video of then United States Vice President Kamala Harris began circulating online. In it, Harris seemed to describe herself as incompetent and a “diversity hire.” The video was eventually discovered to be a deepfake, created using artificial intelligence. Around the same time, Bangladesh was inundated with falsified political content before its general elections, including fake movies targeting opposition personalities and falsely claiming candidates had dropped out of the electoral campaign. After elections in Pakistan, the former Prime Minister Imran Khan, who was in jail at the time, addressed supporters in Pakistan via an AI-generated speech. Meanwhile, Sri Lanka has been rocked by controversy after a fake video of Donald Trump apparently backing one of the presidential candidates emerged, triggering charges of electoral disinformation. These instances are part of a bigger shift in production, dissemination, and consumption of political information.
The advent of AI has not only exacerbated the quantum but also altered the nature of misinformation entirely. Old-fashioned misinformation relies on false statements or twisted narratives, whereas deepfakes utilise one of the most reliable pieces of evidence in contemporary culture: what we see and hear. AI-generated misinformation and disinformation were ranked as one of the most serious global dangers facing societies in the immediate future in the World Economic Forum’s Global Risks Report 2024, and among the threats most likely to precipitate a catastrophic crisis. Microsoft Threat Analysis Centre also noted that national as well as international players are leveraging AI-generated material more and more to influence election outcomes, with elections in nations such as India being identified as prospective targets.
Why India Is Especially Vulnerable
In the Indian context, the threats posed by deepfakes are particularly acute. India is the world’s largest democracy with a huge and incredibly diversified electorate characterised by strong language, religious, regional, caste and socio-economic cleavages. These identities are often the basis for political mobilization. The electorate is particularly responsive to information that taps into social fault lines or reinforces party affiliations. This is made more difficult by the demographic nature of the Indian electorate. The recent election statistics [AV1] show that the elderly are an increasingly major element of the electorate. They comprise around one-sixth of all electors in India, and they exercise tremendous influence because they vote at a very high rate. At the same time, disinformation research has found that senior citizens and those with poorer education or digital competence may have more difficulty in judging the objective reality of online material and are more likely to believe or disseminate misinformation.
India’s Existing Legal Framework
As AI-generated electoral manipulation becomes more sophisticated, India’s regulatory reaction is scattered and primarily reactive. Unlike jurisdictions such as the European Union that have opted for specific AI law in the form of the AI Act[AV2] , India has [AV3] opted for a light-touch regulatory approach uses existing legal frameworks. So, there is no separate law on deepfakes or AI-generated political content. Liability, on the other hand, is scattered over a patchwork of regulations[AV4] such as the Information Technology Act, 2000, the Bharatiya Nyaya Sanhita, 2023, the Copyright Act, 1957 and the Representation of the People Act, 1951, each of which addresses only individual portions of the problem. The Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021, impose due diligence responsibilities on online intermediaries, which have been augmented by a series of advisories issued by the Ministry of Electronics and Information Technology (MeitY). The 2024 AI advisories suggested that intermediaries take more responsibility and be more transparent as well as proactive in deploying artificial intelligence detecting systems, and to remove deepfake content as soon as possible after complaints, following the spread of deepfakes in late 2023. But these actions are still executive advice, not strict statutory responsibilities, and they rely mainly on post-facto content removal, not the prevention or detection of distorted content before it reaches millions of voters. Thus, the current regulatory framework in India is concerned with the repercussions of misinformation generated by AI, rather than the inherent threats provided by AI itself, leaving significant regulatory gaps in respect to electoral integrity, political deepfakes and algorithmically amplified disinformation. [AV5]
Way Forward
India’s fragmented approach is at odds with the global moves to adopt dedicated AI governance. Sections 4 and 5 of the European Union’s AI Act, the first comprehensive AI law in the world, are especially instructive. Article 4 introduces an obligation to promote AI literacy among providers, deployers and affected persons. Regulation alone cannot fight AI generated misinformation; users need to be able to spot and critically engage with AI-generated content. Article 5 meanwhile, prohibits AI practices that pose unacceptable risks to fundamental rights. While not directly legislating on electoral deepfakes, it supports the idea that certain AI applications are inherently in conflict with democratic values.
Other jurisdictions have also recognised the need for dedicated AI governance frameworks. South Korea passed the AI Framework Act, Japan adopted the AI Promotion Act, and China issued its AI Labelling Rules in 2025, each with different regulatory philosophies, but all agreed that AI needs a dedicated framework rather than the application of pre-existing laws. India need not mimic any of these models. But what these developments do is highlight the need to move beyond minimal and scattered statutory provisions and executive advisories and towards a comprehensive AI governance framework. Such a framework should set out clear obligations relating to transparency, accountability, AI literacy, and content labelling. While such legislation would regulate AI across the board, not just electoral content, but it would also inevitably increase the quality of democracy itself by creating more explicit standards for the responsible development and deployment of AI systems, including those that can influence political discourse.
Conclusion
The threat of deepfakes goes well beyond spreading false information. The character of democratic discourse itself is at stake. When voters can no longer trust what they see and hear themselves, the foundation of informed electoral participation is called into question. India’s present laws can deal with some aspects of AI-generated damages, but they are fragmented, reactive and ill-equipped to handle the pace and sophistication of present as well as future technology. As artificial intelligence becomes more and more ingrained in political messaging, regulation must adapt with it. In the era of deepfakes, defending democracy needs more than eliminating manipulated content after it goes viral; it necessitates rebuilding public trust before the next election cycle commences.
About the Author: Prisha Gakhar is an incoming third-year law student at Jindal Global Law School, with a keen interest in economics, IPR and corporate law.
Image Source: Al Jazeera https://share.google/5mLNNZXWUjLFJvyat