The government made it clear on Wednesday that information that uses artificial intelligence to improve clarity, educate, or make technical adjustments without deceiving does not have to be labeled as synthetically generated.
A day after the government tightened regulations for social networking sites like YouTube and X, the clarification was released in the form of frequently asked questions (FAQs). The updated regulations require that illegal content be removed within three hours and that any artificial intelligence (AI)-generated and synthetic content be clearly labeled.
Synthetic Generated Content (SGI): What is it?
The Ministry of Electronics and IT (MeitY) clarified in its frequently asked questions (FAQs) that not all AI-assisted production or editing is considered Synthetic Generated Content (SGI).
“Not all AI-assisted production or editing is considered SGI. According to the ministry, content is only considered SGI if it is produced or modified artificially, algorithmically, or in a way that makes it seem genuine, authentic, or accurate and is probably indistinguishable from a real person or actual event.
As long as they don’t significantly change, distort, or misrepresent the content, context, or meaning of the original material, routine or good-faith operations like editing, formatting, enhancement, technical correction, color adjustment, noise reduction, transcription, or compression won’t be considered SGI.
Technical and educational purposes are exempt.
Content produced in good faith, such as training and instructional materials, presentations, file size reductions for quicker uploads, and posting alerts, will not be regarded as synthetically generated content, the FAQ explained.
Similarly, the SGI category will not include the use of technology to stabilize shaky videos, transcribe interviews, reduce background noise from audio recordings, or adjust color balance.
Additionally, the ministry declared that regular formatting, design, or production of papers, presentations, PDF files, training or instructional materials, or research outputs will not be considered SGI, provided that they do not produce fraudulent documents or electronic records.
Furthermore, as long as they do not materially alter the original content, AI tools that add subtitles, translate speeches without changing the content, create summaries or tags for search optimization, provide audio descriptions for users who are blind or visually impaired, or improve clarity by lowering echo or distortion will be exempt.
When will AI content be considered illegal?
But the government clarified that such content will not be covered by these exclusions if artificial intelligence (AI) techniques are used to create falsified IDs, fraudulent certificates, phony official documents, or created electronic records. Instead, it might be seen as fake records or illegal SGI.
The clarification further stated that, as long as no significant portion of the underlying content is created, changed, or altered, the use of computer resources purely to enhance accessibility, clarity, quality, translation, description, searchability, or discoverability will not be regarded as SGI.
New IT regulations will go into force on February 20, 2026.
The Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021 were amended by a gazette notice published by the Ministry of Electronics and Information Technology. The date of the new regulations’ implementation is February 20, 2026.
The modifications are a reaction to the increasing abuse of artificial intelligence in the production and dissemination of pornographic, misleading, and fraudulent information on social media.
Authorities have noted an increase in artificial intelligence (AI)-generated deepfakes, non-consensual intimate images, and deceptive videos that mimic real-life events or people and frequently spread quickly online.
Mandatory metadata and quicker takedowns
Under the amended rules, platforms must ensure faster takedowns of unlawful content, mandatory labelling of AI-generated material, and embedding of permanent metadata or identifiers with AI content. The regulations also expedite the process of resolving user grievances.
In order to stop the spread and amplification of illegal synthetic content, the updated framework holds social media companies and AI tools accountable.
