India's lawmakers are pivoting from a risk-minimisation approach to an innovation-maximising role in the governance of AI. This is welcome because existing laws - as a GoI panel on AI said this week - provide adequate protections against the new tech, and excessive regulation could hinder its development.
Specific gaps can be plugged across laws as AI risks emerge, with a special emphasis on India-specific risks that need to be evaluated and counteracted. Accountability within the AI ecosystem needs to be enhanced through a graded liability framework.
This is a fair approach in which the impact of the tech is regulated, not the tech itself. It keeps an open mind about the positive effects of technological transformation while ensuring guard rails are in place and offer adequate protection.
Two models are available globally for regulating AI. The first applies rules to specific use cases, which is the Indian position. This governs the development of AI in specific industries through existing and evolving rules for each sector. This builds a scaffold to restrict AI within legal limits across a cross-section of uses. A separate list of no-go areas, such as the use of AI in warfare, can be compiled alongside this.
The second model attempts to regulate the tech at its source through strictures on, for instance, computing power. This is more controversial, with lawmakers and tech developers differing over its merits. It constitutes an extreme risk-mitigation approach, keeping in mind the possibility of AI being able to circumvent some, if not all, use-based regulations. Naturally, the first approach is gaining popularity over the second among lawmakers.
In either case, lawmakers have acquired the capacity to assess risks and can insist on transparency and graded ethical standards from AI developers. As New Delhi prepares to host the fourth in a series of global summits on regulating AI - following previous iterations in Bletchley Park, Seoul and Paris - India is aligning its framework with the evolving consensus.
Specific gaps can be plugged across laws as AI risks emerge, with a special emphasis on India-specific risks that need to be evaluated and counteracted. Accountability within the AI ecosystem needs to be enhanced through a graded liability framework.
This is a fair approach in which the impact of the tech is regulated, not the tech itself. It keeps an open mind about the positive effects of technological transformation while ensuring guard rails are in place and offer adequate protection.
Two models are available globally for regulating AI. The first applies rules to specific use cases, which is the Indian position. This governs the development of AI in specific industries through existing and evolving rules for each sector. This builds a scaffold to restrict AI within legal limits across a cross-section of uses. A separate list of no-go areas, such as the use of AI in warfare, can be compiled alongside this.
The second model attempts to regulate the tech at its source through strictures on, for instance, computing power. This is more controversial, with lawmakers and tech developers differing over its merits. It constitutes an extreme risk-mitigation approach, keeping in mind the possibility of AI being able to circumvent some, if not all, use-based regulations. Naturally, the first approach is gaining popularity over the second among lawmakers.
In either case, lawmakers have acquired the capacity to assess risks and can insist on transparency and graded ethical standards from AI developers. As New Delhi prepares to host the fourth in a series of global summits on regulating AI - following previous iterations in Bletchley Park, Seoul and Paris - India is aligning its framework with the evolving consensus.
