According to the recent statements from the Ministry of Electronics and Information Technology, the government plans to regulate artificial intelligence (AI) “through the prism of user harm”. The increasing popularity of AI with Large Language Models (LLMs) like ChatGPT has reignited discussions on digital privacy and trust, especially in relation to Digital Public Goods (DPGs) and Digital Public Infrastructure (DPI). AI has the potential to enhance quality of life but also raises concerns about privacy and human dignity. This is globally significant, but particularly important for India’s rapid digitalisation.

DPGs are freely available digital resources that promote education, research, innovation, and development for the benefit of society as a whole. They include software applications, educational materials, open data sets, and more. DPIs consist of foundational technological systems, networks, and services that facilitate digital interactions. This includes broadband networks, data centres, cloud computing platforms, digital identity systems, connectivity, storage, and security, etc. DPIs play a crucial role in supporting the delivery of DPGs, public services, and driving digital transformation across sectors like banking, education, healthcare, e-commerce, and governance.

LLMs learn from the digitised wisdom in books, websites, publications, images, and more, and exhibit enough AI to be able to produce human-like responses in natural languages to any complex questions or prompts. If you thought Google Search was impressive, try an LLM. It can answer your most difficult questions or syndicate precise information on any abstract topic, in almost real-time.

Going forward, tight integration of LLMs in all walks of life is inevitable. The vision is not far-fetched when a tribal farmer in the remote corner of the country, will be able to “speak” in their local dialect, on a fine tuned LLM powered phone, and transact complicated business on India’s eGovernment portals or anywhere else in the world.

But what is the guarantee that the farmer would indeed be speaking to an authentic business or government system? And the LLM would not be providing back some fake or misleading responses because it was hijacked by some hacker or biased by some special interest? This is where an endless debate gets ignited. Trustworthiness of LLM-augmented DPIs and DPGs is subjective and depends on multi-dimensional contexts, user experiences, and perceptions, besides transparency in governance, compliance of security and privacy rights, and openness and accessibility of DPIs.

Based on such factors, academia and media often rank countries like the US and Germany relatively higher in terms of digital infrastructure, governance, and open access initiatives. These countries have longer established frameworks and regulations that promote trust in their digital ecosystems. China has also made significant advancements in digital infrastructure, but popular views on its governance and accessibility may differ. Strong public-private partnership can be observed as a foundational trait in all countries with relatively mature DPIs and DPGs.

Indian DPGs and DPIs are increasingly being built using modular and open-source technologies that enable ‘interoperability’, which facilitates the exchange of information between different arms of the public and private sectors, thereby vastly improving the speed and scale of service delivery. The use of open systems allows for better integration into existing setups and helps maintain transparency, further encouraging greater efficiency.

Strong foundations

While the Constitution does not explicitly address digital public goods, its provisions on fundamental rights, directive principles, governance, and judicial review provide strong foundations for promoting trust and enabling the development of DPGs and DPIs in India. The freedom of speech and expression empowers Indian citizens to routinely utilise the social media, blogs and online news portals to voice their opinions, engage in public discourse, and contribute to the digital public sphere. The landmark judgment of the Supreme Court in 2017, in the Justice KS Puttaswamy (Retd) vs Union of India case, recognised the right to privacy as a fundamental right, and ushered a new era of policies and regulations in the country’s digital realm.

The Directive Principles of State Policy in India promote scientific and technological temper. The Digital India campaign since 2015 has transformed the country into a digitally endowed society and knowledge economy, promoting digital literacy, connectivity, and access to digital services. The Right to Information Act, 2005, enables citizens to access government information.

The Indian judiciary through judicial reviews and courts safeguards digital rights and ensures DPGs and DPIs are open to all, non-excludable, and non-rivalrous. India is brimming with innovative initiatives.

During the creation of digital platforms, two approaches that ensure efficiency and trust in the system are: security-by-design and privacy-by-design. While the former makes the systems robust, resilient and hardened against malicious attacks, the later builds trust. Another way to ensure trust in systems is to not make them dependent on only the government. It is imperative that the responsibility of creating, maintaining and delivering DPGs, DPIs and LLM integrations, is shared amongst multiple public-private stakeholders to guarantee checks-n-balances, improve accountability, and reduce risks.

As India seeks to harness the benefits of digital technologies for societal progress, addressing privacy concerns and building trust in DPG utilisation, DPI reliability and LLM accuracy become crucial.

Ensuring user privacy, data protection, information integrity, and transparent governance frameworks are essential for fostering trust and enabling responsible digitalisation in India.

Mittal, IAS, is District Collector of Jashpur, Chhattisgarh, and Jain is Vice President and Chief Security Architect, SAP, US. Views are personal