EU Artificial Intelligence Act: Comparing India’s approach with principles embedded in first of its kind EU legislation

The European Union passed a landmark legislation on Artificial Intelligence on March 13, 2024 with an overwhelming majority with plans to implement the same by the year end.

The said Act is the first such comprehensive framework across the world. It aims to regulate and govern the booming technology that accompanies with it both the fervour and fear of the future. The European Parliament through the first of its kind legislation proposes to protect the diverse aspects of fundamental rights, democracy, the rule of law and environmental sustainability from the impact of high-risk AI while simultaneously encouraging and supporting technological innovation in AI.

The consolidated legislation has bifurcated the technology into four aspects for easy regulation and adherence, that is, prohibited, high-risk, limited risk and minimal risk. The sector dealing with biometric identification in health, law and education are categorized under high risk systems and are destined to meet rigorous requirements, human oversight, security and assessment, whereas the systems interacting with chatbots and image generation programmes are classified under the limited risk category, attracting informed consent of the users.

The legislation has a zero- tolerance policy towards AI models that engage in the manipulation of human behaviour and exploitation of human vulnerabilities such as human race, religion or sexual orientation. However, negligible risk systems such as spam filters and smart appliances are termed as minimal risk and shall be required to adhere to the existing laws. The Act further makes the producers of general-purpose AI systems accountable for furnishing the material used to train their models and to comply with the EU copyright laws.

With the advent of such concise, firm and comprehensive legislation, the impact is preordained to be felt globally, including in India which has already commenced its work on a framework for responsible AI. The EU legislation is already being criticised for its stringent regulations and stifling innovation. India also faces similar challenges, wherein stringent regulations addressing the risks arising in the emerging proliferation of deep fakes are perceived as barriers for promising start-ups and the enhancement of human lives.

The Indian Government has recently addressed the said issue by balancing the risk and regulation through an updated advisory issued by the Ministry of Electronics and IT on March 1, 2024 wherein it has recommended that all intermediaries or platforms ensure that all under-tested or unreliable AI foundational models/ Large Language Models (LLM)/ Generative AI, software or algorithms or further development on such models shall be made available to users in India after cataloguing the probable intrinsic frailty or unreliability of the output generated. The objective behind the advisory is to regulate and make untested AI platforms accountable for deploying any such undependable AI models. With an exception provided to start-ups, the said advisory is an attempt to give way to the transformative possibilities of AI models furthering innovative skills and technological upgradation in the country while balancing the rights of citizens to be informed of any potential hazards.

The sudden upsurge of AI in India - wherein conducive start-up ecosystems, supportive governmental initiatives, rampant adoption and innovation of AI in corporates and educational institutions has led to enhanced competence, accountability and promising innovative processes - also raises potential and concerning issues of ethics, data privacy, transparency, and concerns of biases and fairness fed into the AI models.  However, the Indian ecosystem is addressing these pressing issues by regulating the collection, processing and storage of personal data aligning with International standards, its standardization and certification fostering the components of liability, fairness and innovation.

Congruently, India appears to have adopted a distinct approach to AI regulation wherein it prioritizes harnessing its potential rather than strict regulation, as inferred from the recent statement of the IT minister. This encompasses the intent of the Central government which is contemplating work towards the standardization of responsible AI, as India has already addressed aspects of ethical concerns and risks through its National Strategy for AI report released in June, 2018.

It is incumbent to pay heed to the fact that India has been a step ahead when it comes to shaping its AI regulation and that by implementing The Digital Personal Data Protection Act, 2023 (DPDP) India has already addressed the forthcoming aspects of the processing of digital personal data in a manner that recognises both the rights of individuals and the requirement to process such data for legitimate purposes. The DPDP Act, 2023 covers under its ambit issues of personal data, its processing, the concepts of data processor, digital personal data and artificial juristic person which further incorporates the commercial and technical usage of AI applications. However, with the rise of International Algorithmic Law there is a need to have a global governance field where the development of regulatory tendencies around the humanized use of AI could be aimed towards the global good and developments of democracies across the globe.

Additionally, the proposed Digital India Act, to be rolled out after the 2024 Lok Sabha elections, aims to address the imbalance between digital news publishers and big tech platforms with the larger objective of promoting and fostering an open internet to prevent its monetization by two or three global market players such as Google. The Digital India Act dialogues changed the approach of the Indian government towards achieving the Digital India Goals, 2026 which aims to formulate global standards for cyber laws and not specifically the area of AI as compared to the extant EU Act. The 2026 Digital India Goal is backed by the Atmanirbhar Bharat objective which aims to make India a one trillion dollar economy by 2025-26 and further the cause of global innovation and entrepreneurship and to ensure that India emerges as a significant trusted player in the global value chains for Digital Products, Devices, Platforms and Solutions.

The EU initiative is being referred to as the ‘Brussels Effect’ for global players, including India, wherein the EU has again attempted to unilaterally regulate the global markets by laying down the pathway for stakeholders and global players to adapt and integrate the standards laid down by the EU while framing their domestic laws, which assists access to the EU market with reduced compliance burdens.

While India has already embarked on its journey to work towards the AI for All initiative, it has also co-chaired an India-EU Trade and Technology Council meeting in November, 2023 aimed at strengthening the strategic partnership on trade and technology, inclusive of cooperation on trustworthy AI. Hence, the EU law might act as a reference to Indian policymakers while administering its AI For All campaign, which aims to demystify AI for people from all walks of life and simultaneously acquaint its citizens with AI and its related pros and cons. The recent take of the Indian minster of State for IT further clarified the stance of the Indian government to harness AI and not demonize it. It also showed that the government is working on a model for AI regulation to foster a common protocol for AI when the minister expressed the need for an International Alliance on AI to step beyond abstract principles and approach regulatory harmonization.

Apart from the various compliance burdens that the Indian companies serving EU clients are expected to adhere to, there comes an additional impact on the risk assessment criteria of present AI models against the EU Act’s criteria. The Act comes with the leverage of innovation opportunities, but simultaneously, it also imposes a burden of compliance costs on Indian IT sectors, especially the medium-sized sectors. Thus, it is too early to assess the immediate impact of the EU Act on Indian policies and stakeholders. However, the diverse incidental implications can be deliberated and includes the following:

Global Standards Influence: While the EU has always been the regulatory trendsetter, as companies across the world seeks to comply with the EU laws to access its market of over 500 million consumers, India may consider aligning some of its AI policies with EU standards to ensure trade and collaboration are not hindered.

Potential Collaboration Opportunities: India may consider this opportunity to further its objective of Make in India and Make for World initiatives, wherein joint initiatives, research projects and harmonization of standards between two regions may result in enhanced R&D and responsible development and deployment of AI.

Cross-border data flows: The collaboration may facilitate data exchange and collaboration in AI research and development.

Policy Learning: With the preceding implementation of EU regulations, Indian policymakers may assess its effectiveness and challenges faced by EU member states and adapt such learning to the Indian context.

The concept of AI is a double-edged sword. It carries with it the growth aspect of promoting innovation, R&D and start-ups but is also accompanied by the risks and hazards of potential vulnerability to human lives and rights. AI technology is undergoing a drastic reformation and change in the current times. Hence, its regulation and standardization needs equally equipped legal backing across the globe. While the EU is at the forefront of International Communities working towards the progress and enhancement of AI models, India being the world’s largest democracy and most populous nation cannot stay behind in its approach to tackle the pressing issue of AI, which is developing at breakneck speed, and its related challenges. While the EU initiative is still away from its implementation date, India should aim to work towards international cooperation and coordination among global entities to formulate robust AI governing standards and mechanisms to match the footsteps of EU and stay ahead in its globally recognised human-centric approach embodying the principles of Vasudhaiva Kutumbakam [The world is one family.]. India too needs a robust and comprehensive framework that balances the diverse aspects of the right to innovate and the right to protect. With technology affecting every aspect of human lives, the issues of bias, privacy and the future of humanity are also equally relevant and needs thorough legal backing. The advent of the EU legislation is thus to be considered as a starting point for governance built around technology and could pave the way for Indian policymakers and stakeholders to work towards a comprehensive AI draft in a lightning fast speed, to further the cause of larger interest of mankind and human-centric innovations. Overall, the direct and indirect effects of the EU Act on the Indian marketplace and policies may vary. However, it brings in both challenges and opportunities for India to align its AI strategies with global standards, foster innovation and collaborate with other regions in shaping the future for responsible AI.

About the authors: Shweta Bharti is the Managing Partner of Hammurabi & Solomon Partners. Pranshu Singh is an Associate-II at the Firm.