Introduction
With the intent to steer the growing field of Artificial Intelligence (AI) towards responsible growth and innovation, the Ministry of Electronics and Information Technology (MeitY) issued an advisory on the use of AI models on March 15, 2024 (Advisory)1. This Advisory is a continuation of the previous Advisory dated December 26, 20232 and in supersession of the advisory issued on March 01, 2024 (Former Advisory). The primary focus of this Advisory is twofold i.e., (a) addressing the risks associated with under-tested or unreliable AI models and (b) mitigating the challenges posed by deep fakes. This step emphasizes India's evolving regulatory posture on AI, hinting at a more structured framework likely to shape the future of AI governance.
Why has the MeitY issued the Advisory?
While, at its core, the government acknowledges the exponential growth and integration of AI across various sectors, India does not yet have a dedicated law for AI and, there has been a concerted effort to incrementally introduce guidelines and notifications aimed at AI regulation. This includes legislative / regulatory developments (including advisories) issued under the Information Technology Act (IT Act) and Rules, which now requires online platforms to actively monitor and remove AI-generated deepfakes (post a complaint being reported). The Advisory emerges against this backdrop, driven by several key considerations:
- User Protection: Aims to protect users from the risks associated with deploying under-tested or unreliable AI models.
- Content Regulation: Requires intermediaries to diligently manage content, aligning with the Information Technology (Intermediary) Rules 2021 to restrict unlawful content.
- Accountability: Reinforces the accountability of intermediaries in content management and promotion.
- Prevention of Deepfakes: Addresses the growing concerns around AI-generated deep fakes in India, setting regulations for their creation and dissemination.
- Future Regulatory Direction: Signals a prospective regulatory landscape for AI, with current voluntary compliance possibly paving the way for mandatory
Scope of Applicability
The Advisory extends to all intermediaries and platforms dealing with AI models at various stages of development, including large language models (LLM), generative AI, software(s) and algorithms that are either in testing, beta, or deemed unreliable. While Minister of State for Electronics & IT , Rajeev Chandrasekhar post issuance of the Former Advisory had clarified that it is aimed only at "significant platforms" and excludes start-ups, however, the same has not been formally specified in this Advisory. Unlike the Former Advisory, this Advisory appears to broaden the scope of its applicability beyond "significant platforms" to encompass a wider array of organizations involved in AI development and deployment.
Implications and the Way Forward for Significant Platforms
While the Advisory is not in itself legally binding, it reaffirms obligations under existing laws to further deepening the compliance mechanism. The notable changes brought by the Advisory are:
- Regulation of Content: Intermediaries and platforms utilizing AI must ensure their content does not breach legal norms.
- Electoral Integrity: There's a mandate to prevent computational resources from being used in ways that could bias, discriminate, or spread misinformation affecting electoral processes.
- AI Testing and Deployment: Eased requirement for procuring explicit governmental authorization before deploying under tested AI models, aiming to foster innovation within the Indian AI ecosystem.
- Terms of Service Intermediaries and platforms are required to explicitly inform users of the legal repercussions associated with engaging in unlawful information.
- Labelling for Deepfakes: Platforms generating synthetic content that could propagate misinformation must label / embed such content with unique identifiers / metadata for tracing its origin.
These stipulations necessitate significant platforms to reassess and potentially amend their current practices surrounding the development and deployment of AI models.
Conclusion
The Advisory marks a significant juncture in the government's endeavour to regulate AI technology and sets the tone for forthcoming regulatory initiatives. The Indian government is also planning to introduce the Digital India Act to regulate the potential implications of AI and the ethical and safe usage of AI. The upcoming law is currently in the draft stage and has not been made public.In essence, the Advisory signals the government's conscientious approach to navigating the complexities of AI regulation, striking a balance between innovation and legal compliance in consonance with the global legislative trend for regulating AI technology.
Footnotes
1 Available at: https://www.meity.gov.in/writereaddata/files/Advisory%2015March%202024.pdf (last accessed on April 17, 2024)
2 Available at: https://www.meity.gov.in/writereaddata/files/Eradicated_Advisory_26Dec2023.pdf (last accessed on April 17, 2024)
The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.