Author: Aarti Solanki
In the Indian general elections in 2024, the quantity of deepfake videos in circulation on WhatsApp and Facebook was larger than ever. There were two instances of deep-fake videos – one that portrayed a prominent politician speaking offensively and one that alleged that a candidate had promised illicit advantages through audio clips. These were not obscure examples tucked away on unvisited pages, these were deepfake videos that potentially reached millions of voters, and India lacks the legal framework to deal with these cases of misinformation. It is quite worrying that India does not have laws against these new forms of misinformation.
WHAT IS DEEPFAKE
Deepfakes are essentially manufactured data created using artificial intelligence. This technology enables an AI model to produce videos that show people say or do things they never did, or it can put different words into a politician’s mouth via audio, faces, voices, and even body language can be modified to create convincing content, so much so that in the first watch most viewers find it to be extremely real. While deepfakes are considered tools used for entertainment and learning, their use in politics has made them rather damaging, and so have proved the case in India. The primary issue is simple; when citizens do not know what they are viewing is real or not, an election can never really have free and fair results, as a citizen can never vote knowing what has been depicted might not be the reality.
THE CURRENT LEGAL FRAMEWORK AND ITS NON-EXISTENCE IN INDIA
There are no specific Indian laws to deal with deepfakes in an election, and although there are laws regulating obscenity and sexual content in the IT act, neither applies here. As for forgery and defamation, although it exists under the IPC, it presupposes that a human made a false statement. There is a need to incorporate algorithms as one needs to establish whether the deepfake creator should be responsible for an AI-generated lie. While elections use a model code of conduct that regulates false information, there is a clear lacuna concerning the issue of synthetic media. The DPDP Act does cover privacy concerns, yet not with specific mention of elections and the role that deepfakes can play in sabotaging results. All that happened during 2024 general elections where these videos went rampant was that some statements were released warning the public of such videos, yet beyond that no real action was taken. Platform regulations took precedence in controlling these videos, using the platform’s own regulations rather than any Indian law.
THREE PROBLEMS WITH THE EXISTING RULES
The laws lack sufficient definitions, as they are based on human actions whereas the present misinformation arises from an AI’s output, thereby giving no room for holding anyone liable for a mistake of the AI. The onus is currently on the public to prove that any video has been fabricated as opposed to the person responsible proving it is not a false simulation, and none of the agencies have any expertise or legal backing to prevent these videos from circulating.
WHAT HAPPENED DURING THE 2024 GENERAL ELECTIONS
There were instances in both Tamil Nadu and Punjab of candidates in regional elections being framed through fake deepfakes in terms of accepting bribes, and then also regarding separatists. In the case of those regions of India, there was no proper procedure for investigating and taking appropriate action against these fake videos. All that happened was that the various platform managers were using their own individual regulations on deepfakes that could fail in the specific context of India. The consequences of these deepfakes are varied; though it is not always certain whether votes were actually altered or not, they led to an erosion of trust from the voters towards what was portrayed online, and eventually, it was believed that this was the actual scenario and not a fabrication; it eroded the credibility of the election and its outcome.
WHAT ARE OTHER DEMOCRACIES DOING
The UK passed a legislation whereby all altered forms of media will fall under the Online Safety Bill. Under the DSA in the EU, platforms are required to identify any use of synthetic media. The Taiwan Election Act, passed in March 2023 specifically criminalizes the creation and distribution of deepfakes that impact elections. Several US states have already criminalized deepfakes that mislead voters during elections. What India needs to adopt is perhaps the policy followed by the EU- to ensure transparency and labeling of content rather than banning all deepfakes, and ensure that they do not target election-related misinformation. However, this poses a challenge when one has to differentiate between true parody and malicious impersonation of politicians; a clear distinction must be maintained so that candid deepfakes do not fall under careless use. Most importantly, one must ensure that the country is well-aware and protected from what occurs in the regions rather than the urban centers which are the most obvious victims of deepfakes.
WHAT INDIA SHOULD DO
Definitions are the starting point. The IT rules need a clear definition of deepfakes, explaining that an election deepfake is AI-generated or manipulated digital media, where a person is shown to have said or done something they never said or did and which would reasonably cause an individual to believe it is real. Making the creation and/or dissemination of election deepfakes with the intent to mislead a fresh criminal offense either through an amendment to the IPC or a fresh law is necessary. For election deepfakes that attempt to manipulate the vote, stricter punishments should be given. Accidental or obvious parodies could carry much lighter sentences.
A fresh digital crimes unit should be created under the Election Commission and be tasked with looking into and inquiring into complaints regarding deepfakes and related issues during elections. This unit should be empowered to take down identified deepfakes and pursue criminal charges against their creators. It should also be mandatory for social media platforms operating in India to either flag their manipulated content or inform the authorities.
A rise in deepfakes will lead to more untrustworthy media and the inability of many to distinguish real from fake. This must be countered by increasing efforts by both the state and civic society to inform voters of ways in which to spot them.
THE FREEDOM CONCERN
The laws against political speech should be avoided, especially since electoral laws in India have historically been weaponized. The deepfake law must provide protections for satire and parodies, allow for judicial review of the take-downs, set penalties for government misuse of the law, include safeguards and definitions.
Sunset clauses where the deep-fake law is only applicable during the time of elections and then lapsable afterwards should work to prevent overuse of electoral law. The proportionality question should also be considered while drafting the law to make sure it is enough to deter but not restrictive to genuine innovation.
WHAT WOULD HAPPEN WITHOUT ACTION
If no immediate steps are taken by India, the next election is likely to have even more skillfully crafted and widely disseminated deepfakes. Platforms will continue to take their own decisions on moderation, but this could further erode voters’ confidence. Election integrity, already vulnerable, could be compromised.
The 2024 elections served as a warning. Action is needed in the coming election cycle, if we are to protect the integrity of the electoral process. Democracy depends on people making informed decisions; this will be impossible if artificial media can impersonate anybody and invent speech.
