The criticality of responsible behaviour while using AI cannot be over-emphasized. The world is replete with examples of what unethical behaviour can lead to - we all know it. What is often perplexing is how do we enforce the desired behaviour in the Indian context and through a framework. For emerging technologies that have a very powerful impact, it’s always tricky to define the guardrails. A rapidly shifting landscape and adequate legroom that innovators demand (and rightly so) add to the complexity. Towards simplifying things, NITI Aayog has created a working document which addresses these concerns from various angles.

Let’s try and capture the ideas briefly.

The predominant idea is about context-setting. Risk is not uniform and the variability is dependent on use cases and from the sectoral standpoint. While responsible AI in an autonomous vehicle application is about safety, in the case of credit lending, concerns may be around discrimination and fuzziness in financial inclusivity. Moreover, in the case of highly regulated industries, enforcement bodies are not the same. For instance, in healthcare they are National Medical Commission, NeHA and Drug Controller General, whereas in finance, the regulatory bodies are SEBI, IRDA, etc. And this is why a cookie-cutter approach is fallible.           

The Niti Aayog paper introduces an Oversight body – National Strategy for Artificial Intelligence. This body will play an enabling role under the following heads: 

• Principles for Responsible AI in India have to be continuously managed and updated. There are two parts two this. One is about monitoring and updating the Principles. And, two, the interfacing required with various bodies to design a specific mechanism to help translate principles into practice.   
• Research in technical, legal, policy and societal issues of AI are ongoing. Cross-disciplinary research and studying the on-ground impact can happen with frequent interactions – globally & locally.
• Standards & guidelines form the broad guardrails to clarify what is responsible behaviour.            
• Enabling access to Responsible AI tools & techniques. Data availability, data sharing, and support open technology projects are required.   
• Building awareness by reducing trust-related issues & information asymmetry by sensitizing on ethical problems and the skills required to address them suitably.
• Coordinating with sectoral regulators to identify risks and setting benchmarks accordingly.
• Garner a great deal of international exposure to help relevant ministries in developing cross-border data sharing protocols.

In the UK and Singapore, there are similar independent bodies – Centre for Data Ethics & Innovation and Advisory Council on Ethical Use of AI & Data respectively, comprising experts and influential individuals from many different fields. India’s approach has to be participatory as well and success will be determined how deeply and quickly, the linkages are established. That’s why the composition has to be broad-based with expertise generously borrowed from diverse fields – AI, legal, domain, social sciences, ministries etc.

On paper, this looks fabulous. The real challenge is when the execution mode is on. Having a diverse group brings in richness which would not have been possible otherwise. But operationally, it may be a challenge to align diverse viewpoints leading up to a working document. Secondly, what kind of protection would be accorded to the whistleblowers? As these efforts are continuous, policy-making isn’t going to be easy to address all foreseeable ends.

India can unlock $500 billion worth of economic value in the next five years and be the AI garage as well to 40% of the world. It’s imperative that we get the ethics part right and sustain our advantage as a trusted nation in the digital sphere, particularly in AI. 

The oversight body – National Strategy for Artificial Intelligence isn’t what George Orwell imagined in his famed novella, 1984 but a compass that will help society take the right calls. With great power, comes great responsibility, an adage popularized by Spiderman.