Introduction
In the 21st century, technology has transformed itself into newer possibilities, ushering in a digitally advanced and technologically driven era. The advent of Artificial intelligence has brought a seismic shift in various walks of human existence; however, a closer examination reveals an array of latent issues and challenges. In contemporary times, the rising flow of AI-generated deepfakes during elections has raised the eyebrows of experts and policymakers. The unfettered flow of AI-generated deepfakes significantly impacted the recently held General Elections in India, and when viewed within the larger context of “One Nation One Election”, the existing challenges and issues make a quantum leap. This article seeks to analyse the interplay of AI-generated deepfakes with One Nation One Election and address the pressing issues to demarcate a suitable way forward. This research also becomes highly relevant at a juncture when developments related to the implementation of One Nation One Election are gaining traction.
Influence of AI and Deepfakes on Elections
Usage of AI-generated deepfakes during elections has been one of the most pressing issues faced by the contemporary digital age. There are growing apprehensions about the potential usage of AI-generated deepfakes for sensitising, influencing and further polarising the voters. For instance, the creation of cheap audio fakes, propaganda images, AI parodies and the promotion of misinformation has been the talk of the town in the recently concluded Indian General Elections. Further, the political parties at both national and regional levels actively employed deepfakes and AI to campaign and attract voters through a ‘hyper-personalised’ mechanism. Some of the adverse usages of AI during elections include the creation of fake videos to generate false narratives, voice cloning, resurrection of deceased leaders to garner support for the budding leaders, fake videos showing appeal by Bollywood celebrities for a particular candidate and many more.
Further, at a global level, there is no smooth sailing, and the upcoming elections in various countries might face turbulent waters with the rising usage of AI-generated deepfakes. For instance- the French Elections of 2024 witnessed a bombardment of deepfakes by France’s far-right wing on the burning issues of immigration, attacks, etc. Additionally, the usage of a fake AI-generated video by the Turkish president during a rally has been particularly controversial. Besides, South Korea is also prone to rampant usage of AI and deepfakes, which pose high risks to electoral integrity. These instances exemplify the vast potential of AI-generated deepfakes to have disastrous consequences on electoral integrity.
Interplay of AI and Deepfakes in One Nation One Election
In recent years, one of the most discussed electoral reforms has been the application of the One Nation One Election model in India. Under this model, the voting process for both Lok Sabha and State Legislative Assemblies is expected to be held simultaneously on the same day. It is also suggested that this process need not be conducted in a single day but can also be administered in a phase-wise manner, as may deem fit. Here, it is amply clear that, if implemented, such a model will emerge as a grand carnival of electoral democracy, potentially exacerbating the existing electoral issues. Further, the rising usage of AI-generated deepfakes makes it highly vulnerable to issues and challenges of discrete nature.
Firstly, there are rising worries that during simultaneous elections, the electoral manifestos may overlap, and voters may not be able to differentiate between national and regional issues. This could lead to a scenario where national parties prioritize their national agendas, potentially overshadowing regional parties and their manifestos. . In this scenario, AI chatbots or deepfakes may be used as an instrument to generate emotionally evocative messages or proliferate misinformation against a particular group, community or party, which may lead to the polarisation of voters. Further, the rampant digital illiteracy, currently digital literacy stands at a low 38% of the population, will also provide extra steam to the issue wherein the national parties might leverage misinformation to discredit regional parties, posing a threat to India’s multi-party democracy.
Secondly, conducting elections in a vast and diverse country like India gives rise to a plethora of security complications. It has been noticed that a shortage of security personnel arises even while conducting general elections or state legislative assembly elections.Ironically, there is no specific estimation of the quantum of security personnel that will be required for conducting simultaneous elections, posing a serious question on the maintenance of national security at both external and internal fronts. In such circumstances, AI chatbots and deepfakes may be used by religious extremists, separatists, and non-state actors to further complicate national security by spreading misinformation. In light of the flourishing meme culture, where visual information gets spread at just a click, such content might have a ripple effect, complicating national security.
Thirdly, during the imposition of a nationwide model of conduct for conducting simultaneous elections, the usage of AI and deepfakes may pose significant challenges for the administrative machinery, wherein deepfakes might be used to spread misinformation regarding the administration’s work. For instance, deepfakes could circulate false information aimed at lowering voter turnout, falsely depict administrative failures, or spread fake claims of electronic voting machine (EVM) manipulation. This might complicate the electoral atmosphere and undermine public confidence and trust in the administration.
Fourthly, the usage of deepfake intimate content can be used for the character assassination of vulnerable political figures, especially women. Such insidious manoeuvres can be used to produce fake explicit images and videos of candidates which would encroach upon both private and public spheres. In toto, such instances would not only affect the voters but will make candidates highly susceptible to gross abuse. This concern is especially relevant amid growing worries about the potential targeting of regional parties by national parties during simultaneous elections.
Research Findings and Solutions
One Nation One Election is an ambitious attempt towards streamlining elections and ensuring electoral efficiency. However, such a colossal reformation must be initiated, taking into consideration the systemic challenges posed by the rise of AI and deepfakes. This also becomes particularly concerning when according to a survey, 86% of the Indian population believes that AI and deepfakes may significantly impact the functioning of elections. Further, the Global Risks Report of 2024 has predicted that the collusion between misleading information and unrest in society would arise as a major predicament in several nations facing the ballot in the next two years. It necessitates dealing with the issues and challenges posed by AI and deepfakes on a war footing.
Firstly, it is ironic that despite being in a digital age where every next step is guided by technological faculties, there are no uniform laws to govern the usage of AI in India. The terms like “automated processing tools,” “deepfakes,” etc. have not been defined under any statute, and the current Information Technology Act, 2000, does not provide guidance on the regulation of such AI tools. Consequently, it is crucial to enact specific laws to regulate the usage of AI, especially in the context of elections. For instance, recently, Singapore passed the Elections (Integrity of Online Advertising) Amendment Bill, which prohibited the publication of deepfake content during elections.
Further, such enactments must impose reasonable restrictions to ensure regulatory compliance and should be align with the principles of digital constitutionalism , which provides for the rule of law, separation of powers and protection of fundamental rights in the digital arena. Here, inspiration may also be drawn from the election-centric provisions enshrined under the General Data Protection Regulation (GDPR).
Secondly, the mushrooming of AI-generated deepfakes can be curbed by the establishment of specialised and multi-stakeholder fact-checking units or data analysis units, which would be vested with the task of identifying the misinformation and containing its further dissemination. Such a data analysis unit may be established to specifically curb misinformation pertaining to the elections under the aegis of the Election Commission to ensure its independence and fairness. The establishment of a centralised Data Analysis Unit (DAU) akin to one established by the Misinformation Combat Alliance (MCA), an independent venture by 12 International Fact-Checking Network (IFCN) compliant organisations to counter misinformation, can serve as an effective measure to contain the dissemination of deepfakes. The inspiration can also be drawn from Turkey, where fact-checking units like Teyit have yielded positive results.
Thirdly, digital illiteracy provides lubrication to the swift dissemination of misinformation through deepfakes. Thus, it becomes pertinent that civil society organisations or NGOs must step in to conduct digital literacy campaigns in rural areas, slums, etc. Further, the administration must also conduct seminars to spread digital awareness before or during elections. The rise of digital literacy will help people distinguish between fake and genuine information, thereby acting as a barrier to the further flow of AI-generated misinformation.
Fourthly, a catena of security complications will arise with the proliferation of AI and deepfakes in the situation of simultaneous elections at a pan-national scale. In such a scenario, AI can be utilised as a constructive force to enhance security arrangements both on field and virtual fronts. In the United Kingdom, an AI policing tool is used to flag areas where offences might take place. Such tools can be utilised by security forces to enhance security arrangements at the voting site and other areas which are susceptible to violence and offences. At the same time, fact-checking units, anti-misinformation squads, and similar initiatives can be leveraged to prevent the flow of misinformation during elections.
Fifthly, AI can be highly effective for data management and improving transparency during elections. It can assist in establishing a dedicated and tailored channel of communication with the public as per their personal preferences. The AI could aid in the efficient collection and management of real-time information about elections, thereby enabling the authorities to take prompt decisions. Cumulatively, it would have the effect of bolstering electoral security by making the entire process transparent and accessible.
Conclusion
Elections are the channels through which public representation in national governance is ensured, becoming the sine qua non for the flourishment of democracy. However, the burgeoning technological advancements have left a profound impact on various dimensions of elections. One Nation One Election is a bold step towards refining the electoral machinery in India; however, if implemented without taking into account the vulnerabilities posed by the rising usage of AI and deepfakes, it will be akin to avoiding the big elephant in the room. Thus, it is pertinent to enact specific guidelines that impose reasonable restrictions on the usage of AI within elections and ensure that technological advancements are not hindering the integrity of electoral democracy.
This blog is written by Vishnu Sharma and Tejas Paliwal, 2nd year law student, B.A. LL. B (Hons.), Symbiosis Law School, Pune, Symbiosis International University, Pune.