Govt. Submits Status Report on Deepfakes

Syllabus: GS3/Science & Technology

Context

• Recently, the MeitY submitted a comprehensive status report to the Delhi High Court, addressing the growing concerns surrounding deepfake technology.
  • It highlights the challenges posed by deepfakes, particularly in the context of misinformation, privacy violations, and malicious uses, while proposing actionable recommendations to mitigate these risks.

About Deepfake Technology

• The term ‘deepfake’ originates from ‘deep learning’ and ‘fake’ referring to AI-generated synthetic media that manipulates or replaces real content with fabricated, hyper-realistic counterparts. 
• Deepfake models use generative adversarial networks (GANs), where two AI models — the generator and the discriminator — compete against each other to improve the authenticity of the generated content.

Working of Deepfakes

• Data Collection: The AI is trained on a large dataset of real images, videos, or audio recordings of the target person.
• Feature Learning: The deep learning model learns facial structures, expressions, and speech patterns.
• Synthesis & Manipulation: AI algorithms generate synthetic media that can swap faces, alter expressions, or mimic voices.
• Refinement via Generative Adversarial Networks (GANs): The generated content is refined to improve realism and reduce detectable inconsistencies.

Key Concerns Highlighted in the Status Report

• Lack of Uniform Definition: Stakeholders emphasized the absence of a standardized definition for ‘deepfake’, complicating efforts to regulate and detect such content effectively.
• Targeting Women During Elections: Deepfakes have been increasingly used to target women, especially during state elections, raising serious concerns about privacy and the spread of harmful content.

Other Concerns Surrounding Deepfakes

• Misinformation and Political Manipulation: In India, where social media platforms play a crucial role in political discourse, deepfake videos can be weaponized to create unrest.
• Threat to National Security: Malicious actors can use deepfakes to impersonate government officials, leading to misinformation or even cyber warfare tactics that threaten national security.
• Financial Frauds and Cybercrime: AI-generated deepfake voices have been used to mimic corporate executives, leading to financial fraud.
  • In India’s digital economy, such crimes could severely impact businesses and individuals.
• Violation of Privacy and Defamation: Deepfakes are frequently used to create non-consensual explicit content, disproportionately targeting women.
• Undermining Trust in Media: When realistic fake content circulates widely, it erodes public trust in authentic journalism and evidence-based reporting, affecting democratic processes.

Government Response and Legal Framework

• Information Technology (IT) Act, 2000: It provides a broad framework for cybercrimes but lacks specific provisions addressing deepfake-related offenses.
  • Section 66D: Punishes identity theft and impersonation using digital means.
  • Section 67: Penalizes the publishing of obscene material, which can be used against deepfake pornography.
• Personal Data Protection Bill (PDPB) [Now Digital Personal Data Protection (DPDP) Act, 2023]: It aims to regulate the collection and use of personal data. Misuse of deepfakes involving personal identity could be challenged under this act.
• Intermediary Guidelines & Digital Media Ethics Code (2021): These rules mandate social media platforms to proactively monitor and remove harmful content, including deepfakes, failing which they may lose legal immunity under the IT Act.
• Fact-Checking and AI Detection Initiatives: Platforms like PIB Fact Check have been actively debunking deepfake videos spreading misinformation.
  • Indian start-ups and researchers are developing AI tools to detect and flag deepfake content.
• Global Collaboration: India is collaborating with global tech firms and governments to combat deepfakes through policy discussions and AI research initiatives.

Challenges in Regulation

• Intermediary Liability Frameworks: The report raised concerns about over-reliance on intermediary liability frameworks, which determine the extent to which platforms can be held accountable for content.
• Detection Difficulties: Audio deepfakes, in particular, pose significant challenges for detection, underscoring the need for advanced technological solutions.

Recommendations from the Report

• Mandatory Content Disclosure: The report advocates for regulations requiring AI-generated content to be disclosed and labelled, ensuring transparency and accountability.
• Focus on Malicious Actors: Emphasis was placed on targeting the malicious uses of deepfake technology rather than benign or creative applications.
• Improved Enforcement: Instead of introducing new laws, the report recommends enhancing the capacity of investigative and enforcement agencies to tackle deepfake-related crimes effectively.

Source: IE