LEVERAGING RESPONSIBLE TECH
Working towards the responsible and safe usage of AI systems through risk mitigation frameworks
Context
Artificial Intelligence ('AI') is becoming a dominant force within Indian society and the economy. According to Stanford University's AI Index Report, 2022, India ranks within the top 10 nations across the globe in AI investments, and is the top country (surpassing conventional heavyweights like the US and China), in terms of AI skill penetration.
In 2018, NITI Aayog had published India's national strategy document for the adoption of AI in healthcare, agriculture, education, smart cities, and urban mobility. The Indian state has since used AI in predictive forecasting tools in agriculture, in sophisticated detection algorithms in healthcare, and even for traffic control in urban mobility.
The vision of a more technology-centric society is adequately complemented by a conscious recognition of its risks and the need for adopting mitigation strategies. For instance, in 2021, NITI Aayog published a two-part approach paper on the responsible use of AI. The paper focused on adopting seven ethical principles (like accountability, reliability, safety, etc.) and mechanisms while deploying AI. The Centre for Applied Law & Technology Research ('ALTR') at Vidhi contributed to this paper through extensive research.
ALTR also conducted a detailed three-part study on the use of facial recognition technology ('FRT'), in particular, which identifies and verifies individuals against existing biometric and facial records in digital datasets. In India, a more contentious form of FRT usage is at the state level where local law enforcement agencies use it for monitoring and (targeted) surveillance, and in some cases reportedly even at protest rallies.
ALTR's research specifically examined the legal, ethical, and procedural risks through empirically studying how the Delhi Police used it and the potential risk of religion-based discrimination as a result. Researchers also explored the unregulated as well as opaque involvement of private companies by law enforcement agencies in the deployment of AI, and the unrestricted access such companies were consequently allowed.
Responsible use of facial recognition technology in Digi Yatra
Digi Yatra is an ongoing project spearheaded by the Ministry of Civil Aviation, Government of India, to transform airline travel experience into a seamless, paperless, and contactless process – through creation of a traveller's digital identity using facial recognition technology (FRT) on a smartphone. The digital identity is meant to be used at different checkpoints throughout the airport with the Digi Yatra policy document setting out clear standard operating procedures for each stage.
However, given the risks of inaccuracies, bias, potential discrimination, etc. associated with the use of FRT, legislative, institutional, and regulatory measures safeguarding the user become necessary. ALTR prepared a two-part handbook for NITI Aayog to ensure this.
The handbook provides detailed analyses of the risks associated with Digi Yatra's use of FRT, suggests integration of responsible AI principles, and outlines actionable recommendations.
Impact
- Given that the use of AI, and specifically FRT, is likely to rise exponentially as India chases its 5 trillion USD goal, the existing principles for the responsible use of AI need to be grounded in implementable measures. The handbook prepared by ALTR serves as a larger blueprint for operationalising responsible AI principles in different contexts on a case-by-case basis with the aim of minimising risks and maximising benefits.
- The legislative recommendations in the handbook will help govern and regulate the use of FRT in different states in India. For instance, over 20 states in India are using FRT not only in law enforcement, but even, for example, in COVID-19 protocol implementation. It is crucial that such technologies are deployed through an oversight framework, with clear purpose and scope limitations through legislative or regulatory mandates.
'NITI Aayog has been spearheading the idea of responsible and ethical use of AI within India. Having published principles of responsible AI for India last year, it was felt that the same must be enforced in a test case, by ensuring the responsible and safe use of facial recognition tech in the DigiYatra project. Vidhi's work with us on AI ethics last year, as well as their independent research on the risks posed by unregulated facial recognition was insightful. Consequently, NITI Aayog has commissioned a study from Vidhi's Applied Law and Technology Research team to put together a handbook document for integrating responsible AI principles into the ongoing deployment of facial recognition in DigiYatra. The study will serve as a blueprint for how these principles can practically be adopted and enforced in actual use cases. — Anna Roy, SENIOR ADVISOR, HEAD OF DATA MANAGEMENT AND FRONTIER TECHNOLOGIES, NITI AAYOG