As artificial intelligence reshapes industries and digital finance redraws economic boundaries, the risks beneath our connected world are growing just as fast. At IIIT-H, the Security and Privacy (SyPy) Research Group is working behind the scenes to uncover hidden vulnerabilities, defend emerging technologies, and build the foundations of digital trust. “We live in an online world,” says Prof. Ankit Gangwal, continuing, “Our savings move through digital wallets. Our faces unlock our phones. Our conversations are filtered through machine learning systems that predict what we want before we type it. Every swipe, tap, and transaction depends on layers of invisible code. But what happens when that code is compromised?” Prof. Gangwal’s group, is not just asking exactly that question but working relentlessly to answer it. “To secure the future, we must first understand the vulnerabilities of the present,” he remarks. Security failures rarely announce themselves loudly at first. They hide in edge cases, in overlooked assumptions, in code that “should work.”

In an era where large language models dazzle us with fluency, confident reasoning, and near-human responses, Prof. Manish Shrivastava urges caution by pulling back the curtain on AI’s “illusion of reasoning,” and makes a compelling case for smarter data, smaller models, and a more thoughtful future for AI, especially in the Indian context. Prof. Manish Shrivastava’s research philosophy can be best described with two ‘Rs’: “R for research and R for rabbit holes.” Explaining that there are three types of research, the goal-oriented kind which is focused and socially impactful, the opportunistic kind which jumps into emerging gaps in a field and the exploratory type, driven by intellectual curiosity, Prof. Shrivastava elaborates that most of his work falls into the third category. It’s these rabbit holes that have led him deep into one of today’s most urgent questions: Are large language models (LLMs) actually doing what we think they are? Anybody who is using an large language model (LLM) treats it as an intelligent entity. But for Prof. Shrivastava, it is “facts plus language”.

At the recent Business Standard Manthan Summit in New Delhi, experts underscored the urgent need for India to build its own sovereign foundational AI models and strengthen data sovereignty to reduce dependency on foreign platforms and technologies. The panel included Dr. S K Shukla, Director IIIT-H along with representatives from Mozilla and the World Economic Forum’s Centre for the Fourth Industrial Revolution. Dr. Shukla emphasised that AI should evolve into a digital public infrastructure, serving as a base for sector-specific and organisation-level models, rather than relying on proprietary systems like ChatGPT or Gemini trained on Indian data. The discussions also highlighted India’s demographic and linguistic diversity as a key strength that, when harnessed, can benefit AI solutions across the Global South. Panelists noted that achieving AI sovereignty requires coordinated efforts across policy, research, industry, and skills development, alongside investments in computer infrastructure, multilingual capabilities, and strategic partnerships.