Designing India’s AI Safety Institute, Pg 9

India’s AI Safety Initiative

India announced the launch of an AI Safety Institute (AISI) under the Safe and Trusted Pillar of the IndiaAI Mission. Countries like the UK, US, Singapore, and Japan have already established AISIs to mitigate AI-related risks. India’s AISI will follow a hub-and-spoke model, collaborating with startups, academia, and government departments to ensure inclusivity.

Challenges & Local AI Concerns

AI models face accuracy issues and risks of discrimination in the Indian context. Data gaps exist in Indian AI ecosystems, requiring indigenous tools and linguistically diverse datasets. Startups like Karya are addressing AI biases by creating high-quality Indian language datasets.

Global Collaboration & AI Safety Measures

India’s AISI should engage with global AI governance to adopt best practices. The Bletchley Declaration (signed at the UK AI Safety Summit) provides a framework to tackle cybersecurity risks and disinformation. Establishing a global standardized AI safety taxonomy is crucial to ensure consistent terminology across AI stakeholders.

India’s Role in Global AI Governance

As a leader in the Global South, India can help emerging economies lacking AI safety infrastructure. MeitY-UNESCO collaboration highlights gaps in AI ethics and governance, which India's AISI can address through frameworks on bias mitigation, machine unlearning, and privacy tools. India must establish an international AI model notification system to improve transparency and regulatory coordination.

Way Forward

Develop indigenous AI safety tools and frameworks aligned with India-specific challenges, Ensure interoperability with global AI safety networks. Strengthen digital infrastructure and ethical AI governance to prevent misuse.

Mains Mock Question:

"India’s AI Safety Institute (AISI) aims to balance local AI governance with global collaboration. Discuss the challenges and opportunities in regulating AI safety in India."