India is developing Artificial Intelligence rapidly. In the previous blogposts, we covered the global landscape of AI and the cross-sectoral efforts being taken in India. The current blog post undertakes a deeper analysis of the executive and regulatory approach to developing and deploying AI in India.
The current landscape on AI has been developed through recommendations from various reports and discussion papers from the AI Taskforce (Report of Taskforce on Artificial Intelligence) , NITI Aayog (National Strategy on Artificial Intelligence, Two-Part Discussion Papers on Responsible AI) , and the Ministry of Electronics and Information Technology (MeitY) (Reports of the AI Committees) . Focus is laid on key areas, such as:
i. Establishing Centers of Research and Excellence (CoREs and CoEs) for advancing AI research and development.
ii. Enhancing human resources through skill development and capacity-building programs.
iii. Promoting AI literacy among students and the general population.
iv. Enabling data interoperability to facilitate AI development.
v. Implementing legal and regulatory frameworks for AI.
Additionally, these strategy reports and papers emphasize adherence to seven core principles:
i. Safety and reliability,
iii. Inclusivity and non-discrimination,
iv. Privacy and security,
vi. Accountability, and
vii. Protection and promotion of positive human values.
A. CoREs AND CoEs: A Tale of Success in India
India’s successful establishment of CoREs and CoEs has significantly boosted AI research and innovation. CoEs and CoREs have helped to bridge the gap between innovative ideas and practical implementation by facilitating partnerships between enterprises and AI experts. Collaborations fostered by CoREs and CoEs have led to a notable increase in AI patent filings and research publications. India now ranks fourth globally in machine learning publications and eighth in AI patent applications.  The World Intellectual Property Organization (WIPO) recognizes India’s strengths in AI research, forecasting that its patenting activity in this field will continue to grow, as stated in the “WIPO Technology Trends 2019: Artificial Intelligence” report. 
Patents are a useful measure for inventive activity. However, relying on patents to measure successes in AI must take a cautionary approach. In India, non-resident Indians (NRIs) file more patents than residents, while also receiving more patents. This suggests that a significant portion of AI patent ownership is foreign. Patents grant exclusive rights, potentially hindering smaller companies and startups. Some prefer trade secrets, contrary to India’s AI transparency goals. To address these issues, it’s advisable to establish standardized guidelines for CoEs and CoREs regarding patent licensing in AI Centers of Excellence and Research Excellence to promote innovation without unnecessary hindrance.
Further, there is an urgent need to adopt Free and Open Source Software (FOSS) in publicly funded research, which remains a shortcoming in the current India AI strategy. Using open source and open innovation models can boost inventive activity. Although the Central Government has released a policy on open-source software, implementation is lacking. It is recommended that the government make both source code and server code open source to enable systematic audits and enhance AI transparency. ‘Explainable AI’ alone isn’t enough for transparency; a nuanced analysis of the trade-offs between complex, accurate AI and simpler, human-readable explanations is required. The strategy also lacks transparency in human-AI interactions, such as whether AI should disclose its identity to humans. Encouraging private entities to utilize open-source software and incentivizing research on open-source platforms would enable transparency in understanding the underlying training datasets of AI models.
Solidifying Global Partnerships
While AI publications by Indian researchers have grown, collaborative work with foreign researchers remains limited. Only 15 universities out of numerous institutions have contributed significantly to research publications. Between 2010 and 2019, out of 84,384 AI research papers from Indian authors, just 16% had non-Indian co-authors. Strengthening global partnerships in AI research and publications has become crucial, aligning with SDG 17’s emphasis on enhancing partnerships for sustainable development. Increasing collaboration between Indian and foreign researchers through global partnerships is recommended to boost research quality and impact in AI.
B. Capacity Building and Upskilling
In India, upskilling and capability-building programs have significantly increased, with 69% of organizations focusing more on skill development than in pre-pandemic times. This aligns with the UN SDG 08’s goal of promoting sustainable economic growth and decent work for all. The NASSCOM CEO Survey 2019 showed that 74% of organizations upskilled/reskilled over 20% of their workforce, with digital skills being a top concern. Technologies like AI, big data analytics, and intelligent automation have pushed employee certifications. In collaboration with NASSCOM, MeitY’s FutureSkills PRIME initiative has been a significant success in upskilling IT professionals in AI-related practices. UNESCO has recognized it as a pathway to make India a global talent hub in emerging technologies. NASSCOM’s “FutureSkills” initiative has also reskilled a substantial number of employees in digital talents.
To its credit, the India AI strategy has effectively promoted upskilling and capacity-building programs nationwide, motivating employers to invest in these initiatives. CEO Surveys in 2021 and 2022 affirm that upskilling/reskilling employees remains a key focus in the digital-oriented landscape. 
C. Empowering Students and Citizens with AI Literacy
The Indian AI strategy has acknowledged the lack of AI literacy in the country and introduced initiatives to address this gap, including integrating AI education into school curriculums. Educational institutions like IITs, NITs, and IISC, along with government programs like Responsible AI for Youth and AI for All, have successfully enhanced AI literacy among students and citizens. 
While the focus on theoretical and technical knowledge is commendable, the importance of education on legal and ethical AI issues cannot be overlooked. The current curriculum offered by educational boards like CBSE and ICSE lacks courses that align with a human-centric AI approach. Therefore, it has become crucial to promote equality and inclusivity to prevent biases in AI systems, especially concerning dataset representation. The Indian AI strategy should prioritize equality, inclusivity, and non-discrimination principles to ensure fair treatment for all individuals.
Additionally, education systems should encompass grievance redressal mechanisms, an understanding of ‘Trustworthy AI’ systems, and historical biases. This will contribute to a more inclusive and unbiased AI ecosystem. Privacy and security are non-negotiable. The vast amount of personal data used in AI solutions necessitates ethical training for programmers. Currently, there is a lack of courses on AI’s ethical, moral, and responsible aspects. Educating future developers and users about privacy rights and safety is essential for a human-centric AI environment.
It is recommended that the Government develop an academic curriculum covering ethical and legal issues in AI development and deployment, emphasizing transparency, inclusivity, non-discrimination, security, and privacy principles outlined in the India AI Strategy across all educational initiatives. The Swayam portal which hosts free online courses on AI and ML is a welcomed step in this regard. 
Interoperability in data systems allows multiple systems to exchange and use information seamlessly. This characteristic is crucial for developing AI solutions, particularly in sectors like Healthcare and Agriculture.
In India, AI deployment in healthcare is a significant focus, with initiatives like the Ayushman Bharat Digital Mission (formerly National Digital Health Mission) aiming to digitize health services and ensure data interoperability. This mission emphasizes interoperability as a fundamental component, facilitating data storage, sharing, and accessibility, fostering collaboration between researchers, policymakers, and providers. 
The Ministry of Electronics and Information Technology (MeitY) also introduced the “Draft India Data Accessibility and Use Policy” to ensure uninterrupted data flow between government bodies, businesses, and private entities, with certain exceptions. However, it’s recommended that this policy also align with fair information principles of data protection.
Transparency in data interoperability is a critical concern. The absence of statutory governance leaves substantial control with the Executive. The lack of legal protections for citizens exacerbates these concerns, and India’s AI strategy’s approach leans more towards data-optimization than human-safety. Therefore, it is suggested that initiatives promoting data interoperability should make their methods, technology, and operation standards available to users.
It is also recommended that AI systems working on interoperability should generate privacy and security impact assessment reports for policies and infrastructure accessible to all users. Furthermore, frameworks and guidelines should be established to hold stakeholders accountable in case of breaches, theft, or harm to user data.
E. AI Governance and Legal Framework
A robust legal framework for governing and guiding the Indian AI strategy is the need of the hour. India lacks a comprehensive statutory framework for AI, with the recently enacted (yet to come into force) Digital Personal Data Protection Act, 2023, Information Technology Act, 2000, and the Information Technology (Reasonable Security Practices and Procedures and Sensitive Personal Data or Information) Rules, 2011, serving as the primary regulations for the cyber and data protection spheres. There is no specific law or regulation governing AI-based software or technologies. 
Primarily, it is imperative to promote transparency and reinforcement of positive human values. The Central Government has authorized the use of automated tools and software under certain circumstances for content moderation, as outlined in the Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021.  However, the use of AI-based filtering software raises concerns about false positives, censorship, and errors. Without a clear rights-based framework, these technologies may adversely impact the AI ecosystem. The approach to AI for content moderation needs reconsideration, as AI struggles to understand context nuances, leading to potential issues in distinguishing between harmful and beneficial content. The India AI Strategy should prioritize transparency, protection, and the reinforcement of positive human values in such systems, with corresponding legal and regulatory frameworks.
Additionally, it is crucial to build resilient infrastructure, promote inclusive and sustainable industrialization, and foster innovation. India’s regulatory framework for AI-based technologies is in its conceptualization stage, with efforts by NITI Aayog and technology partners to implement AI strategies in various sectors. It is essential to take into consideration intellectual property laws, particularly patent laws, to facilitate innovation and align with Sustainable Development Goal (SDG) 9 to build resilient infrastructure, promote inclusive and sustainable industrialization, and foster innovation.
Accountability must be inhere in the AI models. Accountability mechanisms for AI stakeholders must be a cornerstone of the India AI strategy. Presently, regulatory tools to ensure accountability, such as discrimination impact assessments and investigations by privacy regulators, are lacking. Questions regarding judicial scrutiny of AI-based technologies also remain unanswered. It is recommended that Indian legislators and authorities take inspiration from Canada’s Bill C-27 and its approach to AI systems, drafting country-specific AI laws.  Bill C-27 aims to establish transparency and accountability standards for AI use by businesses and organizations, including impact assessments and disclosure of information for high-impact AI systems. Additionally, the enactment of the Criminal Procedure (Identification) Act, 2022 remains a concern.  While it aims to improve crime investigation, data collection, storage, and privacy concerns that need addressing, the lack of safeguards against excesses is problematic.
AI has increasingly been deployed towards surveillance initiatives, like NETRA, NATGRID, and the Central Monitoring System (CMS) by the Indian Government. These mechanisms aim to safeguard national defense and security. However, they operate as blanket surveillance measures over citizens’ communications across various means, including telephone networks, the Internet, computers, and facial recognition technology (FRT). The absence of comprehensive data protection rights raises concerns about data breaches and citizens’ rights. There is a need for robust guidelines and statutory protections to ensure these surveillance activities are proportionate and adhere to privacy standards laid down by the Supreme Court.
India’s AI strategy undertakes a multifaceted approach aimed at positioning the country as a global leader in artificial intelligence. It encompasses key methods such as research centers, capacity building, AI literacy, data interoperability, and legal frameworks. While it has made significant strides, there are important considerations and areas for improvement. Establishing CoREs and CoEs has been a notable success, fostering collaborations that have boosted India’s AI research standing. However, there is room to standardize patent licensing guidelines and promote open-source principles more effectively. Global partnerships in AI research are crucial for enhancing research quality and impact. While Indian researchers have increased publications, stronger collaborations with foreign counterparts could further bolster the nation’s AI research capabilities.
Capacity building and upskilling programs have succeeded, aligning with the UN’s sustainable development goals. However, the curriculum should include education on legal and ethical AI issues for a more inclusive and unbiased AI ecosystem. Interoperability and data privacy are pressing concerns. Standardized guidelines for data interoperability and privacy impact assessments are recommended to protect user data and facilitate AI development.
Developing legal and regulatory frameworks for AI is critical. These frameworks must prioritize transparency, accountability, and the protection of human values while ensuring data privacy and security by enacting a comprehensive Data Protection Act. Surveillance initiatives must balance national security with individual rights, adhering to privacy standards and legal protections. With careful implementation and continuous refinement, the Indian AI strategy has the potential to establish India as a global AI leader while upholding ethical standards and safeguarding citizen rights and privacy.
 Report of Task Force on Artificial Intelligence, Task Force on Artificial Intelligence on India’s Economic Transformation, Ministry of Commerce and Industry (2018), https://dpiit.gov.in/sites/default/files/Report_of_Task_Force_on_ArtificialIntelligence_20March2018_2.pdf.
 Responsible AI, NITI Aayog (2021), https://www.niti.gov.in/sites/default/files/2021-02/Responsible-AI-22022021.pdf.
 Artificial Intelligence Committee Reports, Ministry of Electronics and Information Technology, https://www.meity.gov.in/artificial-intelligence-committees-reports#:~:text=Artificial%20Intelligence%20(AI)%20is%20expected,as%20the%20fourth%20industrial%20revolution.
 World Intellectual Property Organisation, WIPO Technology Trends 2019: Artificial Intelligence, https://www.wipo.int/edocs/pubdocs/en/wipo_pub_1055.pdf.
 NASSCOM, 2021 Tech CEO Survey: Industry Outlook, NASSCOM, https://nasscom.in/knowledge-center/publications/2021-nasscom-ceo-survey-tech-industry-outlook; NASSCOM, 2022 Tech CEO Survey: Industry Outlook, NASSCOM, https://nasscom.in/knowledge-center/publications/2022-tech-ceo-survey-industry-outlook.
 The National Education Policy, 2020, Ministry of Human Resource Development, Government of India, 2020, https://www.education.gov.in/sites/upload_files/mhrd/files/NEP_Final_English_0.pdf.
 Course Catalog, Swayam Central, https://swayam.gov.in/explorer?category=Domain_3.
 National Health Authority, Strategy Overview: Making India a Digital Health Nation Enabling Digital Healthcare for All, NITI AAYOG, https://www.niti.gov.in/sites/default/files/2021-09/ndhm_strategy_overview.pdf, para 188.8.131.52.
 Information Technology (Reasonable Security Practices and Procedures and Sensitive Personal Data or Information) Rules, 2011, Ministry of Electronics and Information Technology, Government of India.
 The Information Technology (Intermediary Guidelines and Digital Media Ethics Code), 2021, Ministry of Electronics and Information Technology, Government of India, (2021), Rule 4(4).
 Bill C-27, ‘The Digital Charter Implementation Act, 2022’, available here: https://www.parl.ca/DocumentViewer/en/44-1/bill/C-27/first-reading.
 The Criminal Procedure (Identification) Act, 2022, Act No.11 of 2022, available here: https://egazette.nic.in/WriteReadData/2022/235184.pdf.