Synthetically Generated Information regulation under the Information Technology (Amendment) Rules, 2026

Introduction

The age of the internet has evolved drastically over the decades, which started off with the development of computer programming language to now, with the emergence of artificial intelligence, which uses large language models (LLMs) for its training, machine learning, etc., which are highly capable of generating and creating audio, video, and visuals that closely mimic reality. The most widely talked about and debated topic is the generative model of AI, which produces an output based on the instructions (i.e., prompts) input by the users. The output is produced in various ways depending upon the user’s prompt, such as an image, a diagram, a flowchart, or a specific analysis of documents and texts. The AI will produce this output based on the existing database, and it further keeps tailoring its output based on the user's requirements and personalization, i.e., the AI is trained, firstly, beforehand through a database embedded by the developer; secondly, throughout the user’s requirements.

Now, the question remains as to why AI poses a problem in various industries. Because currently, AI is becoming more advanced in such a short span of time, it can now generate audios based on existing databases, mimic human voices, create and generate ‘deepfake’ videos, and manipulate images, documents, etc. of those events or even persons that never existed. Such as generating/creating a video of a celebrity in a specific event that never happened or depicting false events, all these instances show how dangerous AI can be without proper regulation governing them.

India officially recognized this type of content regulation through the new IT (intermediary guidelines and digital media ethics code) Amendment rules, 2026, notified on 10th Feb 2026 and officially in force from 20th Feb 2026. By specifically defining synthetically generated information (SGI), user awareness regulation, stricter intermediary due diligence and compliance, significant social media intermediary due diligence, etc. These developments pose questions regarding copyright, personality rights, misinformation, privacy violations, intermediary liabilities, and the extent of remedy.

But, before analyzing the new rules, it is important to try and understand the existing legal framework and the initiatives taken by countries across the world.

International response and regulation

Many countries have initiated legal frameworks to regulate the issue and concern surrounding the use of AI. For example, the U.K. took a different approach on regulation of AI, taking on a “pro-innovation approach" rather than a stringent approach. Where it mainly focused on growing the AI sector, improving services and also regulating companies developing AI models, etc.

India’s existing legal framework-prior to 2026 Amendment Rules

Under the IT Act, 2000, Section 79 protects intermediaries, who provide a platform for communication/information dissemination, etc., and will not be liable for the content of users if they exercise due diligence, but this again is conditional in nature. If the platform fails to remove illegal content after knowledge of the same, they are not entitled to the provided protection. The actual knowledge arises through court order or government notification.

Section 69A: Gives the MeiTY the power to block online content/information if it threatens sovereignty, integrity, national security, public order, and the security of the state. One of the examples is when the ministry imposed the blocking order when the BBC documentary controversy was released, which raised allegations regarding the Prime minister's involvement during the 2002 Gujarat riots.

Spreading of “fake, misleading ” information: MeiTY issued the IT (intermediary guidelines and digital media ethics code) Amendment Rules, 2023, which laid down that the intermediaries have an obligation to remove any news that is deemed to be “fake, false, or misleading”, by a fact-checking unit FCU solely composed of the central government. (which was to be the Press Information Bureau, PIB). The Bombay High Court gave a split verdict, with Justice Patel stating that amendment be struck down, reasoning that it is gravely arbitrary in nature, as the words are highly subjective and it is ultra vires the Constitution of India. Currently the matter is pending before the Supreme court and it has put a stay on the government's notification that proposed to establish the FCU.

Under Bharatiya Nyaya Sanhita, 2023: AI misuse could, depending upon the circumstances surrounding the case, attract criminal liability for- impersonation, cheating, extortion, obscenity, etc.

While these laws mainly focused on regulating illegal content, the consequence was limited to removal of information and criminal liability. But these render insufficient for regulating AI generated content as a whole.

The new IT 2026 Rules

The MEITY introduced SGI (synthetically generated information) regulation. This regulates platforms generating AI content and imposes duties such as user awareness, labeling of AI content, watermarking, and due diligence obligations. Especially to address the harms of deepfakes, impersonation, misinformation, synthetic media, etc.

Following rules:

Rule 2(1)(ca) defines “audio, visual, or audiovisual information" as audio, video, photographs, sound recordings, audiovisual etc created using computer resources.

Rule 2(1)(wa)-SGI-content that is artificially, synthetically created/modified using a computer resource that mimics or appears real, indistinguishable from a real person/event.

Exemptions- meaning, even if some content are synthetically generated, they would not under the ambit of SGI for the purpose of IT 2026 rules such as good faith editing/enhancement. without misrepresentation, routine good faith creation/ preparation of documents/ materials, use of tools only for accessibility/ clarity etc.

Rule 3 Intermediary due diligence: they must

• ensure user awareness once every three months in a simple manner about the obligations and consequences of use of the platform.

• label the content: 10% surface coverage in case of visual content, 10% of the starting content in case of audio content;

• If a user creates or uploads unlawful SGI content, the intermediary, upon obtaining actual knowledge, must take action—immediate removal or disabling of content access to the content within 3 hours from receipt of notice/order. This has been reduced from the 36-hour timeline given before the amendment.

• If any user raises a flag about SGI content, the intermediary must take action and dispose of it within 7 days. This is again reduced from 15-day timelines previously provided.

• For sexually explicit content, intermediaries are to remove it within 2 hours of receipt of the complaint. This is again reduced from the 24-hour timeline.

• Other due diligence prevention measures with respect to technical measures, explicit prohibition of high-risk SGI categories, and prevention of tampering of such labeling and watermarks.

The rules address the problems of deepfakes, impersonation, misinformation, and manipulation of content; for example, if AI-generated content claims to be an official document or government statement, etc., this is strictly prohibited.

How do the new rules interact with existing legal rights

As previously mentioned, AI enables the generation of content, be it image, audio, video, etc, which is highly capable of mimicking reality. 

AI-generated content raises questions under the copyright Act, 1957- In which content does copyright subsist, does it also include the AI-generated content/work?

As provided under S. 13 of the Act, copyright subsists in ORIGINAL literary, musical, artistic, cinematograph films and sound recordings.

Moreover, the definition of ‘author’ under section 2(d), means authorship for literary, musical, artistic, photographic, and cinematographic films/sound recordings; the author is who created such work, the composer, the person taking the photograph, and the producer, respectively.

Authorship of computer-generated works:

As provided under S. 2(d)(vi), in relation to any literary, dramatic, musical or artistic work that is computer-generated, the person who causes the work to be created. S. 2(ffc) gives a non-exhaustive definition, which  states that a ‘computer’ includes any electronic or similar device having information processing capabilities. AI systems clearly fall within this category. Hence, AI outputs may qualify as “computer-generated works," but the authorship remains with human. However, Indian courts have never yet ruled or addressed fully whether AI prompts qualify for protection or not.

Furthermore, the phrase person who causes the work to be created, is not defined in the Act. This creates uncertainty in cases where AI itself generates content such as: for eg- if a song is generated by AI- lyrics ‘written’, music composed, artwork, videos, voice cloned by AI. The critical question arises as to who is the author of the work. The possible claims/ contentions are.

1. AI developer- who created the algorithm enabling “creation” can be deemed as they “caused the work to be created”, however, the developer did not create this specific work.

2. The user who is prompting the AI- inputs specific instructions, generates songs, with lyrics written, composed etc.

3. Training database creators who essentially “create” work- AI output production is based on the training data. But AI output could satisfy the originality test if sufficient human input is proved.

Another question pertains regarding the “originality” of the work, which is required for copyright protection. Though AI could qualify, the work in which copyright subsists is the original literary, artistic, musical etc works only. The AI’s output itself is entirely based on training data, existing databases, meaning a collection of various sources of music, musical notes, rhythm etc of existing data, raises great issues on the originality of the work meaning it may not be considered original at all.

Moreover, though AI generated work is a very unique one and an amazing advancement in the growth of technology in the digital era, it raises serious concerns of originality, maybe slowly losing originality over the coming decades and generations. If AI's work is combining bits and pieces and altering or sampling such existing databases to generate a new output, it may not be original as required for copyright protection, meaning it reproduces parts of copyrighted works, issue of infringement definitely arises under S. 51 of the Act

Section 13(3) of the Act states that copyright does not subsist in a cinematograph film or sound recording if it incorporates infringing work.

Loopholes of using AI in “Original” works: persons may use AI for lyric writing, composing, beats composition, video generating, etc and label themselves as authors even though the work is not actually theirs. But, again, the authorship rests with the person causing it to create work, meaning the person will have authorship however, the problem only arises if the AI imitates another singer’s voice i.e., AI copies/extracts existing music parts through training data.

The U.S Copyright office granted protection to the novel called “Zarya of the Dawn”. The office partially cancelled registration and stated that text written by human author and arrangements of images is protected, but AI-generated images are not protected.  Question again remains around the test of originality, creative expression of human input in the AI be considered or the fact that ‘creative work’ was solely output-ed by an AI.

Personality rights: in the case of Titan Industries Ltd. v. Ramkumar Jewellers, the Delhi High court held that the right of publicity is the right to control commercial use of human identity.

The copyright Act does not define the term “personality rights”, and has been left to interpretation by the judiciary over the decades, primarily because the term is not capable of covering the entire concept of subject matters to fall under that category. The judiciary has given meaning to the term as characteristics of an individual like their image in public domain, name, etc of an individual. Protection of such a right, (guaranteed under the ambit of Art. 21 of COI) is an important aspect. Any illegal use of a person’s persona is penalised, like under the IT Act for online identity theft, morphing etc. So, any AI voice cloning or misuse, will violate publicity  rights.

Does SGI under the new 2026 rules try to address the problem of AI generated work?

The SGI primarily addresses the issues limited to circulation of synthetic media via intermediaries, it does not regulate or try to address the authorship of ownership.  Now, for example, if a singer uploads an AI- generated song and music video on any social media platform, main questions arise:

1. Is the content SGI- if the content is AI-generated, it may fall within the board understanding of SGI. however, the classification of such content as ‘SGI’ for the purpose of 2026 rules depends on whether it satisfies falls within the definition  along with the exemption mentioned: good faith editing and use.

2. Is there misrepresentation if the artist presents it as their new song, but the song is fully AI- imitation of other artists, reproduces works of copyrighted works, misleading synthetic media, the distribution and labeling obligations are governed under 2026 rules. The rule specifically targets fake or deceptive content, not the artistic use of AI.

The major challenges in implementation of the SGI

The new rules regulating SGI is a huge step towards addressing the advanced digital harms. Several  challenges may arise in the implementation of the rules itself.

Monitoring and detection:

Identification by intermediaries? of synthetic content at a large scale, as social media intermediaries host enormous volumes of users' content, a manual review is out of the question, hence, the platforms likely rely on automated detection systems, which themselves are AI technologies. But, intermediaries are not obligated to detect every use of SGI. instated it obligates intermediaries to act upon such SGI upon notice:

• From court order or

• An intimation from an authorised government officer

Once the intimation is received, the intermediary must comply with it within 3 hours.

The real question: how does the authorised officer identify that the content is unlawful SGI in the first place? The government cannot manually monitor the entire internet, so this is completely impractical. Then the government may use enforcement via complaints by victims, NGOs, etc, secondly, use automated detections AI tools- basically AI tools being used to detect synthetic media/ AI generated media. This creates a huge gap where legitimate content may be flagged as synthetic or incorrectly detecting/ flagging as SGI, this may ultimately lead to executive overreach, where an authorized officer may issue notice to remove content. And given the extreme 3 hour compliance timeline, intermediaries are unlikely to conduct independent assessment of the content, rather they will merely focus on compliance of notices/orders.

Additional burden on intermediaries: rules impose additional due diligence, like labeling, disclosure requirements of SGI content. While they do enhance transparency, it increases regulatory burden on platforms. Smaller platforms may face difficulty in compliance due to limited technical and financial resources.

Future possibility: this framework may also face constitutional scrutiny in the future. If the obligations that are imposed on intermediaries lead to excessive content modifications, or flagging an AI generated video of criticisms, could curb free speech, and be challenged under Art. 19(1)(a) r/w 19(2). But then again, it is difficult to analyse future outcomes, and can only be determined after deployment of its actual implementation, then analyse the extent of flagging content, notice by authorised officer/ government.

Conclusion

The introduction of synthetically generated information regulation under the IT Amendment Rules 2026 truly shows India’s progression towards more adaptability. The rules mainly deals with defining the scope of  and purpose of SGI, audio, video, visual content, it also laid down intermediary due diligence, specifically including significant social media regulation as well. These try to regulate the content on the platforms, i.e., circulation, rather than addressing issues of authorship, originality of AI generated content. These issues are always present within the domain of laws such as copyright Act, 1957, judicial interpretation of personality rights, etc.

Moreover, the practical implementation of the rules remains. The extreme tight timelines to be complied with by intermediaries, cause to be too stringent. This creates risk of potential executive overreach where incorrect content may be flagged by the authority and may even be misused to curb freedom of expression, and almost control the online content.

Ultimately, the question of whether this will be challenged on grounds of constitutionality remains solely on how it is actually implemented and the extent of removals, the balance of enforcement by authority.

 Author: Akshatha K M, in case of any queries please contact/write back to us via email to chhavi@khuranaandkhurana.com or at  Khurana & Khurana, Advocates and IP Attorney.

References:

1. The Information Technology Act, 2000 [Act No. 21 of 2000], https://www.indiacode.nic.in/bitstream/123456789/13116/1/it_act_2000_updated.pdf 

2. Frequently asked questions on IT (Intermediary Guidelines and Digital Media Ethics Code) Amendment Rules, 2026, GOI, MeiTY. https://www.meity.gov.in/static/uploads/2025/10/065b6deb585441b5ccdf8be42502a49c.pdf 

3. The Copyright Act, 1957 [ACT No. 14 of 1957], https://www.indiacode.nic.in/bitstream/123456789/1367/5/a1957-14.pdf 

4. The Judicially Crafted Shield: India’s Evolving Law of Personality Rights.

https://www.scconline.com/blog/post/2025/12/05/india-personality-rights-evolving-law-celebrity-protection/ 

5. India AI Impact Summit 2026: Landmark Global Declaration and Major AI Investment Commitments. PIB

https://www.pib.gov.in/PressReleasePage.aspx?PRID=2234343&reg=3&lang=1 

6. Europe must hit pause on the AI Act. https://sifted.eu/articles/ai-act-pause-opinion-eu-european-startups 

7. AI regulations around the world- 2026. https://www.mindfoundry.ai/blog/ai-regulations-around-the-world 

8. Challenges to the IT Rules 2023, SCO. https://www.scobserver.in/cases/challenge-to-the-it-rules-2023/ 

9. Titan Industries Ltd. v. Ramkumar Jewellers, 2012 SCC OnLine Del 2382 : (2012) 50 PTC 486.

10. Zarya of the Dawn: How AI is Changing the Landscape of Copyright Protection.

https://jolt.law.harvard.edu/digest/zarya-of-the-dawn-how-ai-is-changing-the-landscape-of-copyright-protection