Innovation Over Tight Control: India Treads Carefully with AI Governance Guidelines

AI is dubbed to be the “fourth industrial revolution” and India is well-positioned to become a crucial player in the global artificial intelligence landscape. A number of things are working in its favour. To begin with, it has a 700 million strong internet userbase that makes it one of the biggest AI markets. Next, government initiatives such as the IndiaAI Mission are playing a big role in boosting the native ecosystem for development of AI and cutting the path towards self-reliance.

What India has been lacking, though, is AI governance – a set of oversight mechanisms that will establish clear guidelines for responsible and effective use of AI. But ahead of the AI summit that India is set to host in 2026, the Ministry of Electronics and Information Technology (MEITY) released a set of guidelines, just a few days back, to serve as a blueprint for AI governance.

The guidelines establish key recommendations on six pillars: infrastructure, capacity building, policy and regulation, risk mitigation, accountability, and institutions. Further, it gives a glimpse of how India means to position itself in the AI age.

The Basic Principle

Pragmatism is the one big message that is emerging from the AI guidelines. Unlike the EU, China or South Korea, the guidelines are not solely focused on regulations. Instead, MEITY’s guidelines indicate that AI governance needs to strike a balance between innovation, equity and regulation while steering away from laying down onerous rules and regulations. This is similar to the US’ pro-innovation, reduced regulation policy for AI.

There’s also an emphasis on transparency, not just in isolated processes, but through the entire AI lifecycle – starting from development to deployment to AI’s impact on end-users. With increased AI penetration, building trust is essential to gaining acceptance, yet the problem is that there is an information black box around how AI works. Taking a cue from the OECD principles of AI transparency and explainability, MEITY’s guidelines seek clear communication around what kind of data and algorithms are specific machine learning models trained on, how are they making decisions in different contexts and the potential biases that could affect the outcomes. Although there are no global standards on how to achieve this transparency, this essentially means creating a trail of data, sharing insights on models, systems, liabilities as well as risks. Another focus that is in line with global regulations is maintaining human oversight to prevent over-reliance on AI systems and to handle ethical complexities that algorithms can't navigate alone.

Action Plan

While acknowledging the need for a separate AI legislation in the future, the guidelines suggest working with existing legislation like the IT Act and the Digital Personal Data Protection Act to deal with issues of safety and privacy.

For now, what the framework specifically proposes is a "whole-of-government" approach for regulation that includes the formation of an AI Governance Group – an interministerial body that will coordinate on newer AI developments and advise the government if new laws need to be considered and drafted. The AIGG is to be assisted by another permanent body - Technology and Policy Expert Body, that will monitor the evolution of AI in different sectors such as healthcare, finance, etc, with the help of industry stakeholders. The underlying rationale is: AI is still taking shape, it is too early to make rigid legislations at this stage.

What This Means for India’s AI Ecosystem

While underscoring the need for responsible and ethical AI, the guidelines encourage voluntary commitments over mandates from industry players through red-teaming exercises, periodic transparency reports, data quality checks, peer reviews, and security assessments.

The keyword here is: voluntary. For India's growing AI ecosystem, this creates room to innovate without facing strict compliance requirements and significant penalties. It creates significant opportunities for Indian startups and established players to build a responsible AI ecosystem through domain-specific risk-assessment protocols developed collaboratively with regulators.

That said, there are clear expectations from AI players in this flexible approach. The framework envisions deeper engagement between industry and government through technical workshops and collaborative discussions, led by relevant regulators and departments. Industry players will be expected to demonstrate their adherence to their commitments to creating transparent and ethical AI initiatives by setting up accessible and effective grievance redressal mechanisms

Concerns Around the Applicability of AI Guidelines

At a time when there’s immense pressure for the need of responsible AI, India’s guidelines on AI governance takes a unique middle-course approach by seeking to regulate AI systems instead of the tech. In many ways, it’s a sharp contrast from the stringent regulations of the EU, Chile or China’s frameworks. But while it acknowledges the risks and the need for regulation, without proper legal framework to anchor it, the guidelines show what responsible AI governance could look like in theory, not in practical implementation.

Another concern is whether self-regulation in tech governance is sufficient to create safe AI systems. While this approach sounds ideal in theory, the track record tells a different story. One need only look at the ongoing struggles to control misinformation and hate speech on social media platforms to see the limitations of industry self-policing. The 2018 Cambridge Analytica scandal (where data from millions of Facebook profiles were harvested without their knowledge) serves as another stark reminder that self-regulation alone cannot guarantee the protection of user data and privacy. Given that AI is increasingly being embedded in high-stakes sectors like healthcare and finance, questions arise on the effectiveness of self-regulation or if there's need for a hybrid approach that combines industry initiatives with enforceable government regulations.