AI doesn’t replace judgment, it complements it

You started your career as a systems engineer at TCS before transitioning into deep academic research at IIM Calcutta and the City University of Hong Kong. Looking back at that shift, what was the specific ‘spark’ that made you realise you wanted to study the behaviour behind technology rather than just the systems themselves?

I moved into academia with a specific interest in exploring how machine learning systems work and understanding their applications across industries. Through my PhD coursework, I gradually came to realise that information technology becomes truly useful only when it is embedded within an organisation and people begin to adopt it. As you rightly pointed out, the behaviour surrounding technology is therefore important and integral to IT as a system. This understanding gradually shaped my research focus towards emerging technologies, their complexities, and the human behaviour that surrounds them.

Having taught and researched across such diverse global environments, what is one ‘blind spot’ you consistently see in MBA students today when they first encounter the complexities of the MBA curriculum?

From my understanding, MBA students are expected to develop a breadth of knowledge along with depth in a few domains. Business school gives you the opportunity to explore various disciplines and build expertise in one or more areas. It is not always wise to choose or focus only on the easier subjects and complete the programme without venturing deeper into more challenging ones. The difficult subjects are precisely what help you build a niche that can differentiate you in the future. Moreover, business school merely lays the foundation upon which you must continue to build towards a stronger, long-term vision. The blind spot, in my view, is a myopic mindset among some students — that is, the failure to think in terms of long-term skill development.

Your work in affective computing suggests that our physical gestures on augmented reality (AR) interfaces can actually reveal our underlying emotions towards a product. As e-commerce moves towards ‘spatial computing’, how do we design these interfaces to be predictive and helpful without making the consumer feel their emotional privacy is being ‘mined’?

At this moment, spatial computing is still a costly affair. Mobile AR is relatively cheaper to implement; however, head-mounted display (HMD)-based AR remains significantly expensive and its adoption continues to be very low. A similar story exists for virtual reality (VR) interfaces. That said, HMD-based AR and VR are far more immersive than mobile AR or desktop VR, and the more immersive the experience, the more natural the behaviour that emerges from people engaging with it. Our sensory experiences from the real world can essentially be mimicked here.

Incorporating interactive and gamified elements within AR and VR experiences can further enrich emotional engagement, predicting which can allow product or service recommendations to be personalised accordingly. However, this comes at the cost of individual privacy. I strongly feel that consumers must be clearly informed prior to the experience about what data is being captured and how it will be used. Consent, in this context, is not a formality but a fundamental design requirement.

You have explored wellness management through the lens of smartphones and wearables; in an era of heightened data sensitivity, what is the primary challenge in building the level of consumer trust required for users to share the deeply personal data that makes these AI tools effective?

I think consumers already share a tremendous amount of personal data without being completely aware of it. For example, GPS data, sensor data from wearables or smartphones, app usage data, social media interactions, and so on. Some of this information may not carry personally identifiable or sensitive details. So, if these available data can be used to genuinely empower individuals towards better self-management, I think consumers will naturally begin trusting your intent.

The challenge here is that design philosophy must shift from data extraction to empowerment. If the interface is designed with transparent intent, where the consumer understands what is being sensed, why it is being sensed, and retains meaningful control over it, personal data becomes a tool of assistance rather than surveillance. The moment the consumer feels the system is working for them rather than studying them, the boundary between helpful and intrusive dissolves naturally.

While digital twins are almost always discussed in the context of heavy manufacturing, your research suggests a much broader scope. Which service-oriented industry, like healthcare or retail, do you believe is currently the most ‘underrated’ in terms of how much it could benefit from this technology?

Apart from product design and manufacturing, my research highlights the applications of digital twins in transportation and healthcare as service industries. In transportation, digital twins can simulate entire traffic networks, predict congestion patterns, and optimise routing infrastructure in real time, thereby reducing both operational costs and environmental impact. In healthcare, digital twins of individual patients can assist clinicians in simulating treatment outcomes before actual intervention, thereby improving diagnostic precision and reducing risk. As these industries deal with high complexity and high-stakes decision-making, digital twin technology offers a powerful layer of predictive intelligence that can fundamentally transform how services are planned, delivered and improved over time.

Having analysed the security of massive infrastructures like Aadhaar, you have seen how vulnerable even the strongest systems can be. In a future dominated by deepfakes, what does ‘trust’ look like for a leader, and how do you teach a student to verify the truth in an algorithmic world?

Fake news, including deepfakes, is not a new phenomenon; societies have always dealt with rumours, propaganda and distorted information even before the emergence of smart and connected technologies. However, recent developments in information technology have significantly increased the scale and speed at which such content spreads. What earlier remained limited in reach can now go viral within minutes due to network effects and digital platforms. Moreover, social media algorithms often function like echo chambers, reinforcing existing beliefs and repeatedly exposing users to similar content, which makes misinformation more persuasive and harder to challenge. Thus, while the underlying issue remains rooted in human behaviour, technology has amplified its impact and made the consequences more widespread and dangerous.

To control and mitigate its effects, a two-pronged approach is required. First, there is a need to strengthen digital laws and regulations to act against those who knowingly create and spread such content to generate panic or disrupt social order. Second, there is a need to educate students and citizens to verify and triangulate information from multiple sources and develop a habit of questioning before believing. This approach recognises that while the problem is not entirely new, its present form requires both stronger institutional mechanisms and greater individual awareness.

There is a persistent fear that generative AI might automate the ‘intuition’ out of leadership; in your GenAI-focused MDPs, which human-centric skill do you emphasise as becoming more indispensable as AI becomes a standard boardroom tool?

There are studies that suggest AI can enhance human ‘intuition’, especially in tasks that involve collaboration between humans and machines. However, to effectively leverage AI, one needs domain expertise to ask the right questions and guide the system towards generating relevant and meaningful responses. In this sense, AI does not substitute human judgment but rather complements it, making the quality of human input even more important.

At the same time, leadership abilities are not rendered redundant by AI; rather, they become more valuable. Better leaders are likely to perform even better when supported by AI in certain tasks. In particular, leadership traits such as critical thinking and interpersonal skills remain indispensable. These are deeply human capabilities that AI, in its current state, cannot replicate, and are unlikely to be fully replaced even in the future.

In a field often seen as cold and algorithmic, how do you personally ensure that the ethical, human-first spirit of ‘Magis’ remains the core ‘operating system’ for your students?

Whenever I discuss technology, I make it a point to also engage with its darker sides: how it affects people, organisational structures, and processes. Beyond the benefits, we examine the challenges that emerge with implementation, including issues of ethics, security and privacy. This naturally leads to a discussion on managing change, where students begin to realise that integrating IT with human systems is often far more complex than the technology itself. The real difficulty lies not in adoption, but in aligning people, processes and incentives with technological change.

Even in the case of AI, which is largely algorithm-driven, a major concern is that organisations tend to treat these systems as black boxes. However, it is important to develop at least a functional understanding of how these systems work in order to interpret and explain the outcomes they produce. This has led to the growing importance of the field of responsible AI (a research area I am personally interested in), where ethics forms a core component. As AI becomes more embedded in decision-making, ensuring transparency, accountability and fairness is not optional but essential.

For a student looking to build a career in emerging tech — whether it is robotics, digital twins or AR — what is the most important non-technical habit or mindset they should be cultivating right now to stay relevant over the next decade?

I have a very simple answer to this: one should be curious and eager to learn. It is important to cultivate a habit of exploring and continuously seeking new knowledge. An MBA programme helps develop this in a structured manner, but the process does not end there. You have to continue learning on your own, especially over the next decade.

Interestingly, generative AI models are improving rapidly by learning from more data and through better techniques. Therefore, if you want to keep up with AI and other technological developments, you need to match this pace by continuously upgrading your knowledge and skills.

Xplore magazine is designed to be a bridge between the academic rigours of XLRI and the high-pressure world of industry leaders. In your view, how can a publication like ours best translate the ‘academic foresight’ of your research into a survival guide for future leaders?

I think magazines like Xplore do a good job of translating academic insights into simple, accessible language for industry leaders. Academia often struggles to balance rigour with relevance, and much of its communication is designed for an academic audience, making it difficult for practitioners to relate to. As a result, valuable research does not always reach those who could apply it in real-world settings.

Professional magazines play an important role in bridging this gap by extending academic research into industry contexts and communicating its practical relevance to corporates. This creates a win-win situation for both academia and industry. There is a need for more content that simplifies research insights and makes them actionable for practitioners. This requires not just translation of language, but also contextualisation of findings into real business problems, use cases and decision-making scenarios. Similarly, industry leaders can introduce more relevant, real-world problems to academia through such magazines. These contemporary industry challenges are likely to engage and motivate academics to work on them.