The Ministry of Electronics and Information Technology (“MeitY”) issued an advisory on March 15, 2024 (“March 15th Advisory”) requiring intermediaries and platforms to ensure due diligence consistent with their obligations under the Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021 (“IT Rules”), framed under the Information Technology Act, 2000 (“IT Act”). Concerned with the intermediaries and platforms not taking adequate measures in tackling the misuse of Artificial Intelligence (“AI”) technologies, MeitY issued an advisory on March 1, 2024 (“March 1st Advisory”) inter-alia advising all intermediaries and platforms to obtain ‘explicit permission’ of the Government of India before using and making available any under-tested / unreliable AI models / Large Language Models (“LLM”) / Generative AI models, software, or algorithms for Indian users. However, after the backlash from the Indian start-up industry, MeitY issued the March 15th Advisory to supersede the March 1st Advisory while reemphasizing the Indian government’s expectation that intermediaries must proactively address concerns regarding the growing misuse of AI.
This note discusses the key highlights of the March 15th Advisory and Kaizen Law’s views on this subject.
March 15th Advisory – Key Themes
| S. No | Highlights | Particulars |
| 1. | Regulation of Unlawful Content | Intermediaries and platforms are required to ensure that the use of AI technology, whether in the form of LLMs, generative AI models, or software/algorithms, on or through their computer resources restricts their users from hosting or publishing any content that could be classified as unlawful under Rule 3(1)(b) of the IT Rules, the IT Act, and any other law in force. |
| 2. | Electoral Process Integrity | Intermediaries and platforms must prevent any bias, discrimination, or threats to the integrity of the electoral process through AI technologies, generative AI, software, or algorithms. |
| 3. | Regulation of Under-Tested AI Models | Under-tested or unreliable AI models, software, or algorithms must be labeled to indicate their potential fallibility or unreliability. Intermediaries and platforms must also implement consent pop-up mechanisms to inform users about the limitations of such AI tools. |
| 4. | Generating User Awareness | Intermediaries and platforms must inform users about the consequences of dealing with unlawful information through their terms of service and user agreements, including warning users about potential access removal, suspension, or termination of usage rights and emphasizing the legal punishment for users under applicable law. |
| 5. | Labelling of Synthetic Information | Intermediaries facilitating the creation, generation, or modification of text, audio, visual, or audio-visual information that users could use to create misinformation or deepfakes must label or embed such content with unique metadata or identifiers to enable the identification of the creator and the first originator. |
| 6. | Consequences for Non-Compliance | Non-compliance with the IT Act and IT Rules may lead to penal consequences for intermediaries, platforms, or users, including prosecution under relevant criminal statutes. |
| 7. | Compliance | All intermediaries and platforms must ensure immediate compliance with these directives. |
Kaizen Law’s views
Scope of March 15th Advisory: The March 15th Advisory covers all kinds of AI tools, such as texts and images that the users may generate through AI technologies. This requires intermediaries/platforms to examine how users may misuse their existing AI technologies in breach of all Indian laws, which presents an onerous obligation. The compliance burden would be significant, particularly for technology companies, such as fintech companies regulated by sectoral regulators like the Reserve Bank of India and the MeitY guidelines. Possibly, in the short term and until Indian regulators streamline their respective regulatory guidance on regulating AI, dual-regulated technology entities must evolve a more nuanced approach for aligning with the varying sets of AI regulations.
Applicability: The March 15th Advisory applies to all intermediaries and platforms. While the term “intermediary” is defined in the IT Rules encompassing entities such as search engines and e-commerce platforms, the expression “platform” lacks any definition under the IT Rules. This lack of clarity makes it imperative for all entities engaged in the AI landscape to evaluate if it may be subsumed within the March 15th Advisory and, if so, put in adequate measures for compliance with the fundamental principles under the advisory, as otherwise regulatory guidance may subsequently emerge clarifying that even smaller tech entities involved in AI generation and creation are covered under this March 15th Advisory.
Areas requiring further clarification: MeitY must clarify the standards of compliance recommended in the March 15th Advisory, such as explaining the accepted labelling and watermarking practices that intermediaries and platforms must adopt to comply with the requirements specified therein. Similarly, MeitY’s guidance on preventing bias or discrimination in AI models is prudent, lest intermediaries formulate internal guidelines for evaluating bias or discrimination within their AI models, resulting in the emergence of different practices and subjectivity in compliances and processes.
Blueprint for formal legislation on AI: Unlike the EU, which released the final draft of its AI Act (EU AI Act) on January 26, 2024, to classify AI systems based on each AI system’s threat, India does not have an overarching legislation addressing AI and its application. Therefore, the March 15th Advisory is significant as it provides insight into the government’s approach to regulating AI, especially as the Digital India Act, expected to replace the IT Act, is still in the works.
Binding Nature: The legal binding of the March 15th Advisory is questionable as it has not been issued under any specific provision of the IT Rules. Further, whether MeitY has residual powers under the IT Act is sub-judice before the Bombay and the Delhi High Courts. Therefore, unless the IT Rules are amended to permit MeitY to issue such advisories governing misuse of AI tools, the March 15 Advisory may not be legally tenuous. Nonetheless, the March 15th Advisory remains a significant indicator of governmental policy intent in AI regulation and must be viewed with seriousness, even if it is non–binding.