Election Commission Steps Up Fight Against Deepfakes: Mandatory AI Labeling for Political Campaigns

Why in News

The Election Commission of India (ECI) issued a detailed advisory on October 25, 2025, directing all political parties to prominently label videos, images, and audio clips that are created or altered using artificial intelligence (AI). This move addresses the rising threat of deepfakes, which the ECI says are "contaminating the level-playing field" in elections by spreading misleading information that looks real. The advisory builds on earlier warnings and aligns closely with recent draft amendments to the Information Technology Rules, 2021, proposed by the Ministry of Electronics and Information Technology (MeitY). It comes amid growing concerns over AI's role in shaping public opinion during campaigns, as seen in past elections where synthetic content disrupted voter trust.

Key Points

1. The ECI's advisory requires political parties to add clear labels like "AI-Generated," "Digitally Enhanced," or "Synthetic Content" to any AI-altered or created media shared in campaigns.
2. Labels must cover at least 10% of the visible area in images and videos, or the first 10% of duration in audio clips, and appear at the top of video screens for easy visibility.
3. Parties must disclose the entity responsible for generating the content in metadata or captions, and keep internal records of all AI materials, including creator details and timestamps, for ECI verification.
4. This follows the ECI's January 2025 directive for basic labeling and a pre-2024 Lok Sabha polls instruction to remove deepfakes within three hours of notice.
5. The advisory echoes MeitY's draft IT Rules amendments from October 22, 2025, which mandate platforms to verify user declarations on synthetic content and display labels, with non-compliance risking loss of legal immunity.
6. Deepfakes are defined in the draft as "information artificially or algorithmically created, generated, modified, or altered using a computer resource in a manner that appears reasonably authentic or true."
7. Earlier incidents, like the January 2025 FIR against Aam Aadmi Party (AAP) for AI-generated images of PM Modi and Amit Shah, highlight enforcement needs, as reported by The Hindu.
8. No infographics were present in the report, but the advisory's structure visually emphasizes "directions" from ECI versus "draft rules" from the ministry, underscoring alignment in labeling standards to prevent electoral misinformation.

Explained

What are deepfakes and why are they a threat to elections?

Deepfakes are realistic-looking videos, images, or audio made with AI that swap faces or voices to show people saying or doing things they never did.

Basic theory: They use tools like generative adversarial networks (GANs), where AI learns from real photos to create fakes that fool viewers, spreading fast on social media.

In elections: They disrupt the "level playing field" by misleading voters, eroding trust, and giving unfair advantages, as the ECI notes; for example, a fake video of a leader making controversial statements could sway opinions without traceability.

Indian context: With 900 million voters and rising internet use, deepfakes amplify misinformation; past cases include 2024 Lok Sabha polls where AI clips targeted candidates, prompting ECI's first guidelines.

What is the ECI's advisory and how does it build on past actions?

The October 25, 2025, advisory mandates proactive labeling to ensure transparency in AI use during campaigns.

Evolution: Pre-2024, ECI required deepfake removal in three hours; January 2025 added basic labels; now, it specifies size, placement, and records to make enforcement stricter.

Purpose: To prevent "unwittingly trapping political stakeholders into incorrect conclusions," as per ECI, aligning with constitutional values of free and fair elections under Article 324.

Enforcement: Parties must comply or face scrutiny; ECI can seek records, building on Model Code of Conduct (MCC) violations.

How do the draft IT Rules amendments support the ECI's efforts?

Proposed on October 22, 2025, these amend the 2021 IT Rules to regulate synthetic content on platforms like X and Instagram.

Key features: Users declare if content is AI-made; platforms use tech tools to verify and label it prominently, or lose "safe harbor" immunity under Section 79 of IT Act.

Link to elections: While broad, they aid ECI by curbing platform-hosted deepfakes; for instance, Meta and X must now watermark 10% of content, mirroring ECI standards.

Broader impact: Addresses global concerns, like EU's AI Act; in India, it ties to Digital Personal Data Protection Act, 2023, for consent in AI use.

What are the challenges in implementing AI labeling for political content?

Technical hurdles: AI evolves fast, making detection hard; free tools create deepfakes easily, and labels can be removed or ignored.

Political resistance: Parties may see it as curbs on creativity; enforcement varies by state, as seen in AAP's 2025 FIR case where delays occurred.

Voter education: Many lack AI literacy; ECI and MeitY need campaigns to spot fakes, like unnatural movements or audio mismatches.

Global lessons: US 2024 elections used watermarks successfully, reducing 30% misinformation; India can adopt similar via partnerships.

How does this fit into India's electoral and digital governance framework?

Electoral side: ECI's powers under Representation of the People Act, 1951, ensure fair polls; AI rules prevent undue influence under MCC.

Digital governance: Part of MeitY's push for ethical AI, including 2023 deepfake guidelines; supports National AI Strategy for trustworthy tech.

Future outlook: With 2026 state polls, full IT Rules could mandate AI audits; experts like NS Nappinai advocate proactive scanning to cut virality by 70%.

Societal benefits: Protects democracy by fostering informed voting, reducing polarization from fakes that exploit emotions.

MCQ Facts

Mains Question