Leading AI firms, Indian innovators make 'New Delhi Frontier AI Impact Commitments': Vaishnaw

New Delhi: Frontier AI companies, along with India's own innovators, have committed to advancing understanding of real-world AI usage to support policies on various issues, including jobs, with multilingual and contextual evaluations under the 'New Delhi Frontier AI Impact Commitments', IT Minister Ashwini Vaishnaw said on Thursday.

Announcing the outcome of India AI Impact Summit here, he said under the 'New Delhi Frontier AI Impact Commitments', "leading frontier AI companies, along with India's own innovators such as Sarvam, Bharatjan, Yani and Soket, have come together to make a set of voluntary commitments that reflect a shared vision for inclusive and responsible AI".

"This initiative positions India at the forefront of building a Global South-led perspective on AI governance, one that balances innovation with equity and real-world impact," the minister said.

Under the pledge, the companies have made "two significant commitments", he added.

First is "advancing understanding of real-world AI usage through anonymised and aggregated insights to support evidence-based policy making on jobs, skills and economic transformation," Vaishnaw said.

"The second is strengthening multilingual and contextual evaluations of AI systems to ensure that AI works effectively across languages, cultures and real-world use cases, especially in the Global South," he added.

Together, these efforts mark an important step towards shaping AI that is not only powerful, but also inclusive, development-oriented and globally relevant, he added.

Under the first commitment, participating organisations recognise that building rich measures of where and how AI is diffusing across the global economy helps to shed light on the future of the work and human-AI collaboration; support the development of evidence-based policymaking in areas such as workforce development and education; enable meaningful comparison over time and help advance adoption and opportunity.

As an outcome of the first commitment, the participants will work to enhance analysis regarding global AI adoption for economic purposes.

Consistent with user privacy commitments, the organisations will publish - by the next AI Summit - statistical insights derived from anonymised, aggregated and taxonomized usage data, either directly or (where relevant) through contributions to international efforts.

The second commitment is aimed at strengthening multilingual and use-case evaluations with participating firms, recognising that cross-lingual support is helpful for democratising AI and aspire to improve Al performance and high-quality experiences for users across the globe.

Furthermore, they recognise the value of partnerships with governments and local ecosystems that have or may develop datasets and expertise that support the evaluation of Al systems for local cultural contexts and use cases.

To this end, the participating organisations will work to advance the evaluation of Al systems for global contexts. They will evaluate multilingual capabilities on a subset of languages and cultural contexts while preserving flexibility on the choice of tools, benchmarks, as well as AI systems prioritised for this work.

Moreover, the participants will collaborate with local ecosystems for the development and application of evaluations for underrepresented languages and cultural contexts.