FROM POLICY TO PRACTICE: IMPLEMENTING EFFECTIVE AI GOVERNANCE STRATEGIES IN INDIA

By Dr. Subi Chaturvedi

India has been a technological and innovative global powerhouse in recent years. With a thriving startup ecosystem and a quickly expanding digital economy, the nation is at the forefront of technological advancement. Artificial Intelligence (AI), one of the many technologies reshaping the modern world, is extremely promising for India. But this promise has its own set of difficulties, especially concerning governance, regulation, and policy.

AI technologies are becoming more and more integrated into different facets of society, such as healthcare, banking, transportation, and education, raising concerns about their ethical use and regulation. What is expected to be a market opportunity worth US$ 1,811.8 billion by 2030 globally, adding US$ 15.7 trillion to the global economy, is also raising concerns regarding its impact on users.

How do we make sure that AI is created and applied in a way that minimizes risks and enhances society? The answer lies in effective AI governance strategies – which require undertaking a journey from policy to practice. Like a true Indian fable, it is replete with its own set of demons, heroes, magical shape-shifting animals, healing plants which patch the wounds, friends, allies and partners who build capacity and coalitions and consortiums who often come together based on agendas, issues and common minimum programs.

There are however many dark dungeons to be explored, and many unsurmountable seas of challenges to be encountered because with great power comes great responsibility as many of our superheroes have vouched and there are many doors which you need to keep knocking till one of them opens or you find a window of direction and you see light at the end of the tunnel. In a world of unknowns, persistence, tenacity, and greater good using tech for the last mile can be our armour where there are webs of lies and deceit galore on what is and what can be, powered by reams of misinformation, disinformation and fake news.

In the words of the renowned poet Ahmad Faraz, this couplet best describes our current engagement with AI which roughly translates into,

‘There is a lot of faith I have on your love, the fear of separating from you is constantly there too.’:

Dil ko teri chahat pe bharosa bhi bahut hai Aur Tuzh se bichad jaane ka darr bhi nahin jaata

दिल को तेरी चाहत पे भरोसा भी बहुत है

और तुझसे बिछड़ जाने का डर भी नहीं जाता

Laying the groundwork

India has made great progress toward creating the framework for AI governance. The National Strategy for Artificial Intelligence, which the Indian government unveiled in 2018, lays out a plan to establish India as a world leader in AI research and development. The aforementioned plan placed significant emphasis on the development of a competent workforce, the promotion of innovation, and the ethical and responsible implementation of AI.

India’s commitment to utilising AI for societal benefit has been further solidified by subsequent efforts like the National AI Mission and the National AI Portal. Transforming these policy declarations into practical actions on the ground, however, continues to be a work in progress with several milestones achieved but requiring consistent upgrades and mitigating newer risks to keep pace with the speed of innovation and rapidly expanding ambit of digital India with new and first-time internet users.

The absence of comprehensive legal frameworks specifically designed for AI technologies is one of the main challenges. While some aspects of AI, like data protection and privacy, may already be covered by current laws, specific guidelines and even legislation are eventually needed to address the particular problems that AI presents, such as algorithmic bias, racism, discriminatory content, copyright issues, responsible innovation, diversity and inclusion, accountability, and transparency.

Entrepreneurs, investors, academicians and policymakers are increasingly struggling to keep up with the quick pace of technological advancement and new trends and use cases. The regulatory environment controlling the use of AI applications must change along with the technology. This calls for a flexible and dynamic approach to AI governance, one that is shaped by continuing discussions between stakeholders, engineers, policymakers, and civil society.

Responsible Development of AI

The development of norms and values is a crucial component of efficient AI governance. If AI systems are not developed and implemented responsibly, they could reinforce or worsen already-existing prejudices and inequities. It is crucial to ensure fairness and equity in AI systems in a nation as diverse as India, with its wide range of cultural, linguistic, and socioeconomic distinctions. This calls for an all-encompassing strategy that integrates inclusion, accountability, transparency, and fairness into the development and application of AI technologies.

For AI to be widely used and accepted, the public’s trust and confidence must be increased. This means demystifying AI systems’ inner workings, promoting an environment of openness and transparency around them, and making sure their application adheres to social norms and values. In this sense, public involvement and awareness-raising initiatives are essential for addressing concerns and dispelling myths about artificial intelligence.

Managing the skills gap is just another essential component of good AI governance. As AI technologies advance, there will be a greater need for qualified experts to create, implement, and oversee these systems. To satisfy this need and provide students with the information and abilities they need to prosper in an AI-driven environment, India’s educational system must change. This involves interdisciplinary cooperation, ethical reasoning, critical thinking, and technical proficiency.

Furthermore, in order to guarantee that AI technologies represent the requirements and viewpoints of all societal segments, and we are able to remove the very high entry barriers towards creating localised Gen AI models with adequate cultural context, initiatives to advance diversity and inclusivity in the AI workforce are crucial along with decentralised models of frugal innovation. A bit like India’s mission to Mars, where we were able to successfully indigenise even rocket science!

This requires proactive measures to overcome barriers to entry and create pathways for underrepresented groups, including women, minorities, academia, research centres, and persons with disabilities, to participate in the AI ecosystem. Institutions of higher learning can play a crucial role in making India the hub of everything AI the Indian way!

A Regulatory Framework for Building Trust in Emerging Technology

India’s first global Framework for Integrity, Security and Trust (FIST) was recently created and launched by InMobi with USI and CyberPeace indicating a colossal shift in policy leadership. Despite India being one of the leaders in tech development and adoption, we have usually waited for international standards and frameworks to be created before adopting them. The FIST framework is a multi- stakeholder initiative to make the Internet more open, accessible, transparent, safe and secure.

The framework was created with the aim of developing a digital landscape that ensures responsibility and accountability at each step of the development and deployment of new and innovative technologies while securing the space for the users. With this framework, India has announced its intent to assume the policy and regulatory leadership of the global digital landscape, ascertaining the rise of the global south. This framework has been the precursor to more specific frameworks for AI safety, consumer interest, and entrepreneurship, among others.

The Gen AI Case

Generative Artificial Intelligence (Gen AI) has an unparalleled opportunity for growth and innovation, making it a revolutionary force that is poised to disrupt sectors in India. According to a recent report, Gen AI has the potential to add up to US$ 1.5 trillion to India’s GDP by 2030.

India’s semiconductor mission is both bold and ambitious in its vision and scope. The Hon. Prime Minister Shri Narendra Modi laid the foundation stone of three semiconductor facilities across Dholera Gujarat, which will be India’s first Fab, CG Power OSAT Facility in Sanand, and TEPL OSAT facility in Morigaon Assam on March 13th, 2024. The aim is to build a vibrant semiconductor and display ecosystem to enable India’s emergence as a global hub for electronics manufacturing and design. The government of India is now offering attractive and competitive schemes for Semiconductor fabs up to 50% of project cost, Display fabs, Compound Semiconductor ATMP, and design-linked incentives. There has never been a better time to make in India.

As technology advances, there is an increasing requirement for efficient governance to control related dangers. The rapid advancement of Generative AI technology prompts worries about false information, skewed results, and ecological issues including higher energy use. Transparency, accountability, and adherence to ethical standards must be given top priority by Indian enterprises across the whole lifecycle of generative AI systems.

Indian organizations need to put strong governance structures in place that include transparent decision- making procedures, well-defined lines of duty, and strict cybersecurity measures in order to handle these difficulties. Moreover, it is imperative to interact proactively with dynamic regulatory frameworks on a national and international level to guarantee adherence to rising standards and best practices.

India can fully utilize Generative AI while reducing risks and establishing itself as a global leader in ethical AI innovation by promoting a culture of responsible AI development and application.

To put it simply, negotiating the terrain of Generative Artificial Intelligence in India necessitates a comprehensive strategy that integrates technological know-how with moral and legal issues. By finding this balance, Indian businesses can harness AI’s transformative potential while avoiding its possible drawbacks, which will eventually spur sustainable growth and innovation across a range of economic sectors.

Best Practices for AI Governance: The Road Ahead

For enterprises to guarantee the responsible and significant deployment of AI technologies, it is imperative to establish an efficient AI governance framework. The following are essential best practices that direct the creation and application of AI governance frameworks:

1. Establishing Internal Governance Structures:

The development of strong internal governance procedures inside enterprises is essential to the success of AI governance. This entails assembling working groups made up of diverse and prominent stakeholders, business executives, and AI specialists. These committees are essential in developing the rules that control AI use in the company. The creation of AI business use cases, the distribution of roles and duties, the upholding of accountability, and the evaluation of results are made possible by internal governance frameworks.

2. Stakeholder Engagement:

Building trust and openness in AI governance requires open communication with all parties involved. Employees, end users, investors, and community members are a few examples of stakeholders. Each stakeholder group should be informed on how AI functions, its intended application, and any potential advantages or disadvantages by organizations. Creating official policies for involving stakeholders aids in creating norms and avenues for communication that are unambiguous.

3. Assessing AI’s Impact on Humans:

AI systems that are well-regulated put people’s privacy and autonomy first, abstaining from bias and discrimination. Risks that need to be recognized and reduced include those caused by biased data sampling techniques, low-quality training data, and a lack of diversity in development teams. Adopting strong risk management techniques guarantees ethical and responsible use of AI.

4. Managing AI Models:

As AI models age, accuracy and performance problems may arise. To avoid model drift, lag, fatigue, and preserve peak performance, rigorous testing, model refreshes, and constant monitoring are necessary. Establishing procedures for continuous model management is necessary for organizations to guarantee AI systems function well and produce accurate results.

5. Addressing Data Governance and Security:

Given the sensitive nature of the data involved, data governance and security are critical issues in AI governance. To protect customer data and preserve the integrity of AI system results, organizations need to put strong data security measures in place and follow applicable data privacy laws. Responsible AI deployment is ensured by reducing the risks of data breaches and misuse through the development of AI- specific data governance and security policies.

6. Higher Focus on Corporate Governance:

A lot of entities in the space and key players are startups, often driven with a growth mindset to the exclusion of purpose. The pressure for numbers, and profit over people or planet is quite real. In an environment where there is a funding winter the temptation to make a quick buck and take the project live with unrealistic investor expectations, fear of missing out from competitors is plaguing giant US tech as well as legacy businesses. More than ever it’s important to slow down, build consciously and build to last. What we do today with AI literally shapes the discourse and wipes out millions in value overnight.

7. It’s an arms race:

Keeping AI at an arm’s length is equally important. Much like Indian homemakers or an experienced chef who knows exactly when to bring the curry to boil, add curd or cream without the gravy separating AI infusions across process both in government and the private sector need to paced right and need based. An unhealthy obsession with AI only because of the fear of missing out shouldn’t also be encouraged with an adequate transition time while we build capacity.

8. Fail early fall forward:

For any technological innovation to reach its potential, it is important to create governance frameworks that are light touch, encouraging and facilitative. Once the technology has reached its logical zenith, regulation is a must but over regulation in the early stages can hinder and even discourage innovation. So regulatory sandboxes which allow experimentation will help us to quickly understand what works.

9. Equity over equality:

AI for all is not a mere slogan, rather a pertinent warning. AI has the potential to bridge the gap between the digitally endowed and the digitally limited as well as increasing it. It is important to ensure that we use equitable distribution of resources, focus and effort of deploying AI and training people in it, rather than equal distribution. A neat example would be the digital trickle-down effect of new technology adoption in cities where people end up using one tool or the other just because everybody else is using it while in rural areas, the users are never exposed to new tools and thus never end up using them or learning how to use them. Allowing for decentralised models with a plug and play system and choice-based adoption will go a long way in ensuring everyone has a voice and we do not leave people behind.

10. #HarGharAI:

AI is probably the biggest technological innovation since the fortunate marriage between affordable smartphones and cheap data tariffs. While this led to remarkable digital and economic growth in the next decade, the AI promise is much bigger. While the former gives people access, the latter promises to give them power. AI is transforming industries as well leading to revolutions in many sectors. Take Nvidea for example, which crossed US$1 trillion in value, fuelled by the Gen AI boom as they developed the semiconductors and GPUs to process it. Other components used in developing, deploying or sustaining AI will witness similar growth and success and that is what makes AI so transformative.

Moving from Policy to Practice:

A multifaceted strategy that takes into account the regulatory, ethical, educational, and socioeconomic aspects of AI governance is needed to move from policy to practice. Even though India has laid a strong foundation for AI governance, there is still more to be done to make sure that AI technologies are created and applied in a way that benefits society as a whole.

Government, business, academia, and civil society can work together to unlock and harness AI’s revolutionary potential while avoiding some of its possible hazards and pitfalls in India. Realizing AI’s potential as a positive force in India and abroad would require coordinated efforts so that we can create a truly connected and empowered Digital India where technology becomes a force for good.

About the Author:

Dr Subi Chaturvedi, is a distinguished public policy professional, an AI tech policy expert, and a former member of the United Nations Internet Governance Forum, MAG. She is currently Global SVP, Chief Corporate Affairs & Public Policy Officer, InMobi and FICCI Chair Women in Technology, Policy & Leadership.

Tweets @subichaturvedi
Email: subichaturvedi@gmail.com; subi.chaturvedi@inmobi.com