The Indian government is considering amendments to the Information Technology Rules, 2021, mandating large social media platforms (Significant Social Media Intermediaries or SSMIs) to label AI-generated media.
The proposed rules aim to address the rise of indistinguishable digital alterations, including those used to create misleading content, such as deepfake videos of public figures.
The draft rules propose labels covering at least 10% of the visual area of synthetic videos or 10% of the initial duration of synthetic audio.
Social media platforms are struggling to accurately label AI content, with audits showing failure rates as high as 70% across major platforms.
Detailed Insights:
The current framework for regulating synthetic media is considered insufficient, necessitating amendments to the IT Rules, 2021.
Synthetic media is defined as content artificially or algorithmically created or modified to appear authentic, but challenges exist in differentiating between benign and harmful content.
A tiered-labeling system distinguishing between 'fully AI-generated', 'AI-assisted', and 'AI-altered' content may be more effective than a one-size-fits-all approach.
The proposed rules should be broadened to include creators directly, especially those with a significant follower base, mandating disclosure of AI use in content creation.
Independent information verifiers and auditors can play a crucial role in identifying harmful, fraudulent, and misleading content, supplementing automated detection systems.
Key Concepts Involved:
Synthetic Media: Content that is artificially or algorithmically created, modified, or generated to appear authentic.
Significant Social Media Intermediaries (SSMIs): Large social media platforms with a significant user base, subject to specific regulations.
Deepfakes: Synthetic media, often videos, where a person's likeness is digitally manipulated to depict them doing or saying things they did not.