The spirit of democracy has long been a debated topic in India. With a multitude of perspectives, often conflicting, on the interpretation of the spirit of democracy, it is difficult to characterize it in a single definition. Yet, despite all these conflicting opinions, something that most would agree upon is that electoral integrity remains one of the most essential factors to uphold such spirit. Dr. B.R. Ambedkar famously claimed a parliamentary democracy to mean ‘one man, one vote’, reinforcing fair elections as the very foundation to a democracy.[1] Despite such nature of elections, Indian elections have historically faced many obstacles. These obstacles have ranged from “criminalization of politics, booth capturing, tampering of electoral rolls, intimidation of voters, communalization of elections to the most recent electoral bonds scandal”.[2]
The most recent manifestation of these obstacles with the potential of having a more widespread impact than any of its predecessors is that of the rise in use of generative artificial intelligence in political undertakings with emphasis on the use of deepfake during elections.[3] Picture that it’s just a few days until election day and a deepfake video of your constituency’s leader is released in which he is seen indulging in corrupt practices. The video, although synthetic, would immediately cause widespread chaos in terms of distrust within existing leaders, lack of confidence of the public in political institutions and even low voter turnouts. Even though it may be established soon that the video had been synthetic, significant damage would have been caused already.
Additionally, other minor instances like proliferation of deepfake videos of political leaders making false claims or engaging in inappropriate activities or inherently baseless conversation might incite public distaste and might alter the outcome of the elections. Thus, generative AI content can prove to be highly impactful on the electoral integrity of India. Despite this, the incorporation of AI in the context of politics cannot be discarded without examining its merits.
Proliferation of generative AI content can also be a very effective tool of questioning and criticizing the government for its shortcomings as it is often used to draw up satirical and parodic work that highlights failures with sarcastic jabs.[4] So where exactly do we draw the line between satirical and parodic content protected under freedom of speech and expression and the misleading generative AI content that has potential to be detrimental to a free and fair election.[5] A solution to this can be found in a strong regulatory framework that correctly identifies between inauthentic content and illegal content. This distinction is extremely vital in order to regulate deepfake and generative AI content during elections while also rejecting censorship by maintaining the freedom of speech promised by the constitution. We shall look into these regulatory requirements after understanding the problem and its relevance in contemporary India.
The introduction of artificial intelligence to Indian politics is changing how political parties run their campaigns and devise electoral strategies. The Indian general elections of 2024 saw an escalation in the use of AI in political campaigning as political parties spent nearly $16 billion on campaigning initiatives.[6] This surge came with an increased use of deepfakes and generative AI content aimed at eliminating competition and pushing agendas to accumulate votes. Although a very effective tool when used wisely and legally, there is a very thin line between acceptable generative content and content that creates disorder in an electoral ecosystem.
The Indian general elections of 2024 saw various incidents which crossed such lines, threatening the sanctity of an election. A major incident was when Meta, otherwise one of the most accountable and biggest players in the media industry, allowed 14 AI- generated electoral advertisements that incited violence. These advertisements, all uploaded between 8th May 2024 and 13th May 2024 just a month prior to the general elections, called for hate crimes against Muslim communities and spread rumors based upon religious conspiracies.[7] This shortcoming occurred despite Meta’s claim to carefully identify and remove harmful AI generated content ahead of the elections in 2024.[8]
Additionally, there were also incidents that were directly aimed at the leadership of major political parties. During the elections, a synthetic video of Prime Minister Narendra Modi was released, showing him dancing to a Bollywood song.[9] Additionally, a synthetic video of a prominent leader, Ms. Mamta Banerjee was shared in which she was dancing on an audio of her criticizing her colleague for joining the BJP.[10] The state police subsequently initiated an inquiry, stating that the content could potentially “affect law and order”.[11]
Although all these incidents were a result of synthetic content, there is an evident difference between the situations we have noted. The Meta advertisements inciting violence and Prime Minister Modi’s dance video, for example, are two very contrasting uses of deepfake technology. The Meta advertisements posed a serious threat, potentially inciting violent uprisings by manipulating certain sections of society. On the other hand, the dance video of Narendra Modi was clearly a satirical jab which even the Prime Minister himself found humorous, sharing it on his X account.[12] Would treating both these pieces of content as equally detrimental suffice for the purpose of regulating AI in the electoral framework? There is no simple answer to this. But there is an objective which one can work around.
The basic formula for maintaining fair regulation of AI in politics is limiting the content that affects the fairness of an election, while also upholding the freedom of speech and expression as guaranteed under Article 19(1)(a) of the Constitution.[13] To follow such formula, a flexible regulatory framework is required which can clearly identify between inauthentic and illegal generative AI content.
Deepfake and generative AI content that clearly constitute parody, satire or constructive criticism should be encouraged as they enable an essential attribute of a democracy of having the ability to question a government.[14] Such content, although unoriginal and manufactured, carries no ill intention of misleading people and bringing unfair alterations to the electoral process. It does not violate the sanctity of a fair election and can be deemed as merely ‘inauthentic content’. On the contrary, synthetic content that under the guise of free speech, incites violence during elections, creates disorder, spreads harmful narratives, defames leaders and misleads the citizens of the country crosses the threshold of mere satirical and parodic work and rather constitutes ‘illegal content’ that is detrimental to elections.
The simple difference between inauthentic and illegal generative AI content can be understood through an illustration. Consider the earlier example of a deepfake video of a prominent leader being spread before the general election within which the said leader is seen engaging in corruption, propagated to the public under the impression that it is completely authentic. This would manipulate the masses and create a false narrative against the said leader, thus constituting illegal content. Now consider the same situation, but with a slight difference. The difference being that this leader is widely known to be corrupt with substantial proof of such practices. Picture that the deepfake this time is a parody video in which the leader is engaging in corruption, transitioning into him addressing the electorate directly, claiming corruption to be the most important part of his manifesto.
This second piece of content is clearly a satirical jab at an already corrupt leader, criticizing him for his actions. Such content, although synthetic, cannot be said to be illegal on two grounds. The first being that the video is clearly not claiming to be authentic, evident from the nature of the video. The video’s transition to the leader addressing the electorate clearly establishes its satirical nature. The second is that the video is not spreading any misinformation. It is merely a criticism of the actions of the leader for which he has been incriminated before as well.
The intention behind these two videos is clearly distinct and thus treating them with the same approach would be inherently flawed in principle. Thus, for adequate regulation of deepfake and generative AI content during elections, it is important that we clearly distinguish between content that is merely inauthentic and content that is illegal. We shall now further explore why current legislation on this matter falls short and the amendments that are required.
Elections form the essence of any democracy as they serve as the primary tool through which a citizen can truly realize their power in a representative form of a government. Subsequently, Indian elections are regulated by a comprehensive structure that aims at ensuring fairness.[15] The major constitutional provisions that constitute this framework include Articles 324, 326 and 329 which respectively empower the Election Commission of India (ECI) to oversee and conduct elections, guarantee adult suffrage and restrict judicial interference in elections.[16] Additionally, the process of elections is mainly regulated by the Representation of the People Acts of 1950 and 1951 and the Election Symbols (Reservation and Allotment) Order, 1968[17] which provide structure to electoral administration and its operative functions.
The ECI, established by Article 324, is tasked with supervising, directing and controlling the conduct of elections.[18] In order to facilitate this, the ECI formulates several policies, mandates and requirements for different political actors and parties to comply with. ECI’s major mandate, the Model Code of Conduct (MCC) monitors the actions of political actors during electoral processes.[19] Although these provisions and the Model Code of Conduct have been effective parameters for electoral regulation in India so far, the present framework is not adequate to deal with threats of the digital age, like deepfake and generative AI.
AI regulation in electoral processes is limited to interpreting the IT Act and the MCC or other notifications in a manner which would seem fit corresponding to the situation. This current framework falls short of addressing the scale of problems caused in an electoral system by tools like deepfake and generative AI. This inadequacy has led to the Election Commission having to release specific advisories related to the matter. One such advisory issued was on 24th October 2025, directed at political parties, asking them to comply with some directions, titled “Advisory on responsible use and disclosure of synthetically generated information and AI-generated content during elections”.[20] Although these directives were quite robust, they were merely directed at political parties involved in elections. They did not create any accountability for private individuals that might proliferate such harmful content. Furthermore, such advisories are not largely enforceable as they do not carry significant binding power.[21]
This clearly identifies a need for an authority which not only mandates such provisions on political parties but also regulates synthetic content created by private actors. The authority currently in place for this purpose, the Information Technology Act, 2000 is a broad legislation which regulates technology and AI at a larger level.[22] It does not specifically cater to the regulation of AI use within the electoral framework, thus leaving room for a high volume of subjectivity and interpretational errors. Such subjectivity and errors often lead to what is known as over regulation, which in the context of media is one of the biggest threats to the conduct of a fair election.
The Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Amendment Rules, 2026[23] categorizes deepfake or generative AI content under the ambit of the term “synthetically generated information”.[24] Further it imposes restrictions on the accommodation of any such synthetically generated information that misrepresents a person or real-world and exhibits deception.[25] Although a very conclusive and robust provision in the larger context on generative AI, it does not cater to the technology in aspects of its application in the political context. Additionally, it only creates obligations on the intermediaries facilitating such content and does not instill a sense of accountability in political parties and individual actors.
The reason why this provision is not exactly suitable for an electoral scenario is embedded in the very nature of politics. This broad definition could be manipulated to use against content that provides constructive criticism and questions the work of political parties. As a result, such unspecific provisions being interpreted for the context of elections will create a threat of censorship, being fatal for a fair election. It is well established in both Indian and international jurisprudence that public figures involved in governance can be met with greater criticism than civilians and consequently law will consider the same”.[26]
In the case of Raghav Chadha v. Ashok Kumar and Ors[27], the Delhi High Court recognized that actions by politicians will usually invite harsh criticism which might often be showcased as satirical humor.[28] The court also established that although it does not endorse the use of AI generated deepfake videos to harm the dignity of an individual, it recognizes that AI has become a tool for expression across platforms, which is evident in the political realm and until strong legislation is established to deal with such cases, it is the responsibility of the judiciary to fairly examine the same.[29] Furthermore, international jurisprudence has also reinforced such ideas. In Kulis’ v. Poland, it was held that a higher threshold for criticism should exist for people who voluntarily open themselves to the masses.[30]
Thus, this difference that has been clearly established in criticism aimed at a private individual and a political figure calls for a different approach at treating deepfake and generative AI content aimed at politicians. It is vital to facilitate fair elections while also upholding constitutional guarantees that allow the people of a country to criticize and question their government. For this purpose, it is necessary to introduce a regulation that strikes a perfect balance among unprecedented freedom and over regulation of such content. We can consider certain legislation that exists at the global level, to further improve our understanding of what is required.
The increasing significance of artificial intelligence in numerous fields has garnered a lot of attention worldwide. Owing to such attention, many countries have recognized the need to create a framework that monitors the use of AI in polity. We can delve into laws in the EU, South Korea and Singapore to understand how different nations have addressed the involvement of AI in politics. The Artificial Intelligence Act of the European Union, applying to 27 of its member nations is a legislation that specifically caters to applications of AI in various fields to ensure safe adoption. The act lays down certain provisions for the adoption of AI in the political realm and during elections. Although it does not provide detailed legislation, it gives certain foundational principles that can be interpreted in impactful ways.
Initially, the act, with the objective of addressing risks of undue external interference with voting rights lays down that AI models that can be used to influence electoral results must be categorized as high risk systems.[31] Establishing such a classification allows for clear identification of possible threats and prevents incrimination of lawful use of generative AI, like for the purposes of satire, parody and criticism. Additionally, Article 50(4) of the act mandates the creators of any AI infrastructure to make it known when a piece of media has been synthetically altered if the said media generates any image, video or audio that could be considered a deepfake.[32] The act more importantly establishes that if any media is evidently artistic, creative or satirical, the disclosure mandates would be limited to a manner in which the viewing experience of the same is preserved.[33] Thus, it becomes evident that the EU AI Act is a legislation that aims at a fair balance in regulation of generative AI and deepfake while also maintaining freedom of speech and allowing people access to satirical and parodic works.
Now we can look at some different approaches where regulation has been attempted for the use of synthetic content in electoral campaigns. Singapore’s Elections (Integrity of Online Advertising) (Amendment) Act 2024[34] is a legislation that accounts for the proliferation of generative content during elections. Section 61, added through the amendment act levies numerous restrictions on election advertising and even lays out penalties for the same. The provision establishes that a digitally generated election advertisement that contains an audio, video or audiovisual representation of a candidate engaging in any act that the candidate has not actually done and such representation is realistic enough that members of the general public would reasonably believe it to be real, then the publisher of such content would be held accountable for committing an offence.[35]
Similar to Singapore, South Korea also very recently amended its Public Official Election Act in 2023, realizing the importance of AI governance in elections. But, as opposed to the laws we have discussed earlier, the Public Official Election Act adopts an approach to the AI problem that is a lot more restrictive than its global counterparts. Article 82-8, added through the amendment prohibits anyone from producing, editing, distributing, screening or posting any deepfake videos with the purpose of electoral campaigning from 90 days before election day.[36] Rather than creating a distinction between content that can pose threats to an election and content that is protected under free speech, this act eliminates the possibility of any content being generated around the election period. As deliberated upon before, in this generation, AI and AI generated content has become a significant means of expression and curbing its fair use can amount to censorship.
On comparing these three legislations from around the world, we can conclude that they vary highly on the amount of leverage they provide. EU’s AI Act is optimal among these three to draw inspiration from as it provides for regulation with room for interpretation. Singapore’s Elections (Integrity of Online Advertising) (Amendment) Act 2024 although not as flexible as the AI Act, does provide scope for interpretational leverage. The South Korean Public Official Election Act, on the contrary, fails to establish safeguards against censorship as it entirely prohibits the use of generative AI for a large period before the elections.
The three legislations we have analyzed above can be used as examples of how institutions are beginning to understand the importance of regulation of AI within electoral frameworks. Although the degrees at which these legislations operate are varying, they establish the same idea: AI when used without any safeguards, poses a major threat to the essence of a democracy by manipulating and creating disorder in the electoral framework. It is indispensable to recognize this threat and establish a robust legislation that aims at regulating such generative AI, without threatening the rights of a citizen to freedom of speech and expression under Article 19(1)(a) of the Constitution of India.[37]
The legislation we have discussed the need of should function on the basis of certain principles that reinforce the goal of adequate AI regulation in elections while also upholding the inviolability of free speech. For such objectives, the legislation should be built on 5 foundational principles. These principles are: (i) Mandatory Disclosure of Synthetic Content; (ii) Metadata Preservation; (iii) Algorithmic Audits Compliance; (iv) Inauthenticity and Illegality Distinction and (v) Safeguarding Freedom of Speech and Expression. These principles when drafted, applied and interpreted effectively will provide a robust structure for regulation of AI in India’s electoral framework.
Mandatory Disclosure of Synthetic Content: Any political parties, private individuals, media houses or any intermediaries involved in circulation of deepfake or generative AI content which is intended to be part of a campaign or directed towards any political actor must be obligated to visually label such content with textual or graphical cues. Studies indicate that AI labelling is extremely effective in altering the believability of a piece of artificially generated content.[38] This obligation must exist regardless of the distinction between inauthentic and illegal content as it would act as a primary requirement in AI regulation, curbing misinformation and creating accountability on the producer for any deepfake or generative AI content with respect to the narrative such content creates.
Metadata Preservation: Any publishers of deepfake or generative AI content must be prohibited from hiding any embedded metadata that their files must contain. Such metadata gives additional information regarding digital content and can amplify the ease of tracking the origins of such content. Metadata, upon examination, can provide authorities with crucial information like date, time, device and location of creation of content.[39] Such embedded details can act as a tool to increase the accountability of the content publisher and can make the tracking of harmful content much more efficient.
Algorithmic Audits Compliance: Developers of AI systems that are widely used must comply with algorithmic audit requirements on a regular basis. International legislation like the Artificial Intelligence Act requires the programmers of AI infrastructures classified under high levels of risk to engage in algorithmic audits and carry out “evaluations, assessments of systemic risks and finding ways to mitigate them and tests to comply with transparency requirements”.[40] Regular audit compliance would ensure safe usage patterns of AI and would flag any high-risk parameters.
Inauthenticity and Illegality Distinction: We have earlier established our concept of differentiating deepfake and generative AI content into the categories of inauthentic and illegal. When we use the term “inauthentic content”, we refer to content that although is synthetically produced, does not mislead the masses into believing something that is untrue. On the other hand, “illegal content”, while already being inauthentic, attempts to establish fabricated narratives as believable in order to manipulate the masses. The question arises how one would exactly identify the distinction between the two. A few characteristics of the piece of media can be taken into consideration for this. Firstly, it is very clear that if a piece of artificially generated content merely depicts an incident that has already occurred or a fact that is proven to be true and uses such a thing to criticize an individual or political party, it would not amount to being illegal. Additionally, If the content in any media makes it obviously clear that it could not possibly be real and has been synthetically produced, it would not be reasonable to consider it illegal under the guise of misrepresentation. Furthermore, any generative content that is intended to be a parody or satire and carefully complies with all other requirements of labelling, metadata preservation while also refraining from other infringements and blasphemy, should not be immediately categorized as illegal without careful examination. This inauthentic and illegal content distinction can become an effective tool to establish the balance between regulation and free speech.[41]
Safeguarding Freedom of Speech and Expression: Free speech lies at the core of the Indian Constitution. Since India’s independence, this principle has empowered its people to criticize and question the government with the objective of bringing out optimal productivity for the country. Satire, parody and sarcasm have always been powerful tools for this purpose and now with the inception of generative AI, this tool can be used more effectively than ever before. During the 2010’s, ‘So Sorry’, a satirical animation famously aired by news channel ‘Aaj Tak’ often targeted famous politicians for their shortcomings and criticized their hypocrisy. This show gained immense popularity amongst children and effectively increased their political acumen, awareness and interest, all while highlighting the government’s failures.[42] Thus, with changing technology, it is important that we embrace their use for political discourse. Rather than restricting AI’s use in politics for the threats it creates, emphasis should be put in attempting to ensure fair use and distinguish clearly among such fair use and the potential threats.
Thus, India’s existing framework governing the deployment of deepfake and generative artificial intelligence content in election processes falls short of striking a fair balance between effective threat management and safeguarding free speech protected by the Constitution. The combined interpretation of ECI’s Model Code of Conduct and IT Act 2000 has proven to be an insufficient means of regulation of AI which has been evident through instances like the ECI’s release of advisories during elections and the increased use of problematic AI patterns during the 2024 elections. According to the 2026 Global Risks Report, India is at a high risk for the spread of misinformation and disinformation. The report also recognizes the risk of AI’s impact on elections in countries with such risk.[43] It is evident that a need for a dedicated legislation on AI regulation in elections has arisen, with countries like Singapore, South Korea and EU member states already taking steps in the same direction. It is vital for a multi-party democratic system like India to understand the urgency of this matter and establish a framework that effectively caters to its diverse political system. While establishing the immense need for threat management, it has also been established that such management must be done in a way that upholds the Indian Constitution’s promise of freedom of speech and expression. AI regulation should never be embodied as censorship and effective laws and authorities need to be established to ensure this. Compliance frameworks that include AI labelling, metadata preservation, effective disclosures and algorithmic audits should be mandated while the authorities themselves also comply with establishing clear distinctions between inauthentic and illegal content with the purpose of upholding free speech. Thus, in order to allow India’s people to put forth their perceptions on the spirit of a democracy, we must first allow them access to free, fair and protected elections that form the very foundation of such a democracy.
[1] Reghunath LG, ‘The Doctor’s Wise Words on 25 November 1949’ (2023) Supreme Court Observer.
[2] Sia Shruti, ‘Election Reforms in India and Challenges before the Election Commission’ (Manupatra Articles, 24 August 2012) <https://articles.manupatra.com/article-details/Election-Reforms-in-India-and-Challenges-before-the-Election-Commission> accessed 25 July 2026.
[3] Raluca Csernatoni, ‘Can Democracy Survive the Disruptive Power of AI?’ (Carnegie Endowment for International Peace, 18 December 2024) <https://carnegieendowment.org/research/2024/12/can-democracy-survive-the-disruptive-power-of-ai> accessed 25 July 2026.
[4] Nora Nahae Kim, ‘Deepfakes and Democracy: Free Speech vs. Election Integrity’ (2025) Columbia Science and Technology Law Review.
[5] Raghav Kohli, ‘Expressive Conduct and Article 19(1)(a) of the Indian Constitution: A Purposivist Approach’ (2022) 17(1) Asian Journal of Comparative Law 41–68.
[6] Mithil Aggarwal and Janice Mackey Frayer, ‘How AI Is Changing Elections Around the World’ (NBC News, 6 June 2024) <https://www.nbcnews.com/news/world/india-ai-changing-elections-world-rcna154838> accessed 25 July 2026.
[7] Waquar Hasan, ‘Meta Approves Ads Inciting Hate Against Indian Muslims During Elections: Report’ Clarion India (15 May 2024) <https://clarionindia.net/meta-approves-ads-inciting-hate-against-indian-muslims-during-elections-report/> accessed 25 July 2026.
[8] Ekō and India Civil Watch International, Meta Fails to Stop Violent and Inflammatory AI-Generated Ads Targeting Indian Voters (Ekō 2024) <https://aks3.eko.org/pdf/Meta_AI_ads_investigation.pdf> accessed 25 July 2026.
[9] Aditya Kalra, ‘Dance Videos of Modi, Rival Turn Up AI Heat in India Election’ (Reuters, 16 May 2024) <https://www.reuters.com/world/india/dance-videos-modi-rival-turn-up-ai-heat-india-election-2024-05-16/> accessed 25 July 2026.
[10] ‘Lok Sabha Election 2024: AI Videos of PM Modi, Mamata Banerjee Heighten Misinformation Risks’ Hindustan Times (16 May 2024) <https://widgets.hindustantimes.com/india-news/lok-sabha-election-2024-ai-videos-of-pm-modi-mamata-banerjee-heighten-misinformation-risks-101715852241029.html> accessed 25 July 2026.
[11] Ibid.
[12] Ibid.
[13] The Constitution of India 1950, art 19(1)(a).
[14] UNESCO, Guidelines for the Governance of Digital Platforms: Safeguarding Freedom of Expression and Access to Information Through a Multistakeholder Approach (UNESCO Publishing 2023) <https://unesdoc.unesco.org/ark:/48223/pf0000387339> accessed 25 July 2026.
[15] Prithviraj S and Shilpa ML, ‘The Legal Regulations Governing Elections in India- Addressing the Functional Issues’ (2025) 2(7) International Journal for Legal Research & Analysis.
[16] Ibid.
[17] Ibid.
[18] Election Commission of India, ‘About ECI’ (Election Commission of India) <https://www.eci.gov.in/about-eci> accessed 25 July 2026.
[19] Ibid.
[20] Election Commission of India, Advisory on Responsible Use and Disclosure of Synthetically Generated Information and AI-Generated Content During Elections (ECI Directive No 4/Misc/2025/SDR/Vol.XX, 24 October 2025).
[21] Press Information Bureau, ‘Violation of the Model Code of Conduct’ (Ministry of Law & Justice, 8 May 2012) <https://www.pib.gov.in/newsite/erelcontent.aspx?relid=83354> accessed 25 July 2026.
[22] IndiaAI, ‘Report on AI Governance Guidelines Development’ (IndiaAI, 6 January 2025) <https://indiaai.gov.in/article/report-on-ai-governance-guidelines-development> accessed 25 July 2026.
[23] Sara Mishra, ‘A Case for a Human-Centric AI Legislative Framework in India’ (SCC Online Blog, 26 June 2026) <https://www.scconline.com/blog/post/2026/06/26/human-centric-ai-legislative-framework-india/>accessed 25 July 2026.
[24] Ministry of Electronics and Information Technology, Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Amendment Rules, 2026, GSR 120(E), Gazette of India, Extraordinary, Pt II, s 3(i), 10 February 2026, r 3(1)(b).
[25] Ibid.
[26] ‘Media and Elections’ (ACE Electoral Knowledge Network) <https://aceproject.org/ace-en/topics/me/onePage> accessed 25 July 2026.
[27] Raghav Chadha v Ashok Kumar John Doe and Ors CS(OS) 466/2026 (Delhi HC, 1 July 2026).
[28] Ibid.
[29] Ibid.
[30] Laila Clarke and others, ‘Criticism Against Politicians and its Protection in the Jurisprudence of the European Court of Human Rights’ (OxJournal, 24 August 2023) <https://www.oxjournal.org/criticism-against-politicians-and-its-protection-in-the-jurisprudence-of-the-european-court-of-human-rights/> accessed 25 July 2026.
[31] Regulation (EU) 2024/1689 of the European Parliament and of the Council of 13 June 2024 laying down harmonised rules on artificial intelligence and amending Regulations (EC) No 300/2008, (EU) No 167/2013, (EU) No 168/2013, (EU) 2018/858, (EU) 2018/1139 and (EU) 2019/2144 and Directives 2014/90/EU, (EU) 2016/797 and (EU) 2020/1828 (Artificial Intelligence Act) [2024] OJ L 2024/1689.
[32] Artificial Intelligence Act, art 50(4).
[33] Ibid.
[34] Elections (Integrity of Online Advertising) (Amendment) Act 2024.
[35] Elections (Integrity of Online Advertising) (Amendment) Act 2024, s 61.
[36] Public Official Election Act (Republic of Korea) art 82-8.
[37] Raghav Kohli, ‘Expressive Conduct and Article 19(1)(a) of the Indian Constitution: A Purposivist Approach’ (2022) 17(1) Asian Journal of Comparative Law 41–68.
[38] Chloe Wittenberg and others, ‘Labeling AI-Generated Media Online’ (2025) 4(6) PNAS Nexus 170.
[39] Collate, ‘Metadata in AI Systems’ (Collate Learning Center) <https://www.getcollate.io/learning-center/metadata-in-ai-systems> accessed 25 July 2026.
[40] Warren Z, ‘What the EU’s AI Act Means for Service Firm Professionals’ (Thomson Reuters Institute, 11 December 2023) <https://www.thomsonreuters.com/en-us/posts/technology/eu-ai-act/>accessed 25 July 2026.
[41] Nishtha Gupta and Netheena Mathews, ‘India’s Experiments With AI in the 2024 Elections: The Good, The Bad & The In-between’ (Tech Policy Press, 25 September 2024) <https://www.techpolicy.press/indias-experiments-with-ai-in-the-2024-elections-the-good-the-bad-the-inbetween/> accessed 25 July 2026.
[42] Chitranshi Raizada, ‘Analyzing So Sorry Politoons in the Context of India’s 2024 Lok Sabha Election’ (2024) Journal of Journalism and Mass Communication.
[43] World Economic Forum, The Global Risks Report 2026 (21st edn, World Economic Forum 2026).