Robot dog debacle raises tough questions

THE ongoing mega AI Impact Summit and exhibition in New Delhi was supposed to showcase India's rise as a formidable player in the emerging global AI landscape. But it has ended up being a mirror to everything that is wrong with the R&D and innovation scenario in the country.The parading of a robot dog imported from China by a private university, Galgotias, in its enthusiasm to project itself as an AI innovation hub, has ended in a shameful fiasco.University officials, in media interviews, claimed that the robot was developed as part of a Rs 350-crore AI initiative of the university. Social media posts debunked the claim, adding that the robot was Chinese-made. The claim featured in DD news as well as a tweet of the IT minister. Eventually, the university was asked to vacate the premises, and the minister's tweet deleted.Negative headlines in India and abroad apart, the episode points to the slow drift being witnessed in scientific and academic research in the past several years. India cannot hope to become a global leader in AI and other emerging technologies without addressing the issue.The Economic Survey presented in Parliament last month expressed concern over R&D expenditure intensity despite progress made in this sector. India's Gross Expenditure on R&D (GERD) as a percentage of GDP stands at an abysmal 0.64%, which is substantially below the global average. It is far below that of the leading economies — the US (3.48%,) China (2.43%) and South Korea (4.91%).The spread of research funding is skewed in favour of national laboratories and top institutions like IITs, with the contribution of the private industry and universities lagging far behind. The full impact of the new initiatives, such as Anusandhan National Research Foundation and the Research, Development and Innovation Fund, is yet to unfold.For India to make a mark in any emerging field, including AI, fundamental research cannot be ignored. In recent years, China has raced ahead in research papers and patent filings related to AI and associated fields. The robust university-research-industry combo in China is helping it do so.On the other hand, India is not paying adequate attention to the quality of teaching and research output or industry collaboration in universities, resulting in the menace of plagiarism, fake research, promotion of pseudoscience, retracted papers and other ills. The hype and overemphasis on startups have resulted in our institutions making tall or fake claims, as Galgotias has done.The AI landscape is fast changing. From the novelty of large language models (LLMs) and popular generative AI applications, we are beginning to see real-life impacts of technology. The San Francisco-based AI firm, Anthropic, caused tremors in technology markets as well as stock markets last week with the release of its Claude Opus series of AI tools designed to handle complex financial analysis and long-form coding.The unveiling of the automation tools triggered a panic selloff of the shares of traditional software and legal-tech companies as investors feared AI agents could disrupt the labour-intensive IT services sector. Anthropic has begun selling AI agents to large Indian outsourcing firms, which could lead to large-scale layoffs.India, with this summit, is making a bid to claim its space in the global race by seeking to 'democratise AI' as it fears that AI is getting concentrated in a few corporations and geographies. The idea is to deploy AI across sectors such as education, health, governance, manufacturing, agriculture, etc.Unlike the West, where private capital leads in AI, India has taken the socialist path. AI development is state-funded through the AI Mission and public investments in such initiatives as BharatGen. India is developing 'sovereign' AI models, not commercial ones like ChatGPT. Developing niche models suitable for multi-lingual and multi-cultural environments and deploying them for public good is the objective.Translating the vision into reality, however, will be more challenging. The biggest challenge is the gap in manpower skilled in AI and other new technologies. India has a large IT workforce of about six million people, employed in the IT services sector.A large part of this pool is threatened by agentic AI tools like Claude. While some people will lose their jobs or will have to be retrained and reskilled, we also need a large workforce to develop LLMs and applications as well as safety and regulatory aspects. In such a situation, it is surprising to see overenthusiastic policymakers claiming India will become a supplier of AI engineers to global firms. This is what we did with IT services in the 1990s, which resulted in India lagging in product development. We should not tread the same path with AI.Another challenge is a severe shortage of the high-end hardware required to develop and train LLMs. To develop a sovereign AI model, the AI Mission is aggressively working to establish a GPU cluster, but the high cost of hardware and global supply chain delays are significant hurdles.China is investing massively in creating AI infrastructure to develop a sovereign AI model. In 2025 alone, Chinese spending for creating AI infrastructure was close to $98 billion. Comparatively, India will spend $1.25 billion on its sovereign models over five years, till 2029.The focus should be on developing a robust and sustainable AI infrastructure and reskilling the workforce. Instead, policymakers are promoting captive GPU infrastructure in the form of private clouds owned by global giants (Microsoft, Amazon, Google, etc).This is another form of outsourcing, and more harmful, as large data farms are power and water guzzlers. Giving them tax holidays at the cost of environment is a dangerous idea. Data centres owned and operated by global firms in India do not contribute to the development of Indian capabilities in AI, as is being projected by the government.If the government is interested in promoting research and innovation, it cannot afford to neglect the basics — support to both fundamental and applied R&D and enforcement of academic and quality parameters.Policymakers and ministers should stop making tall claims and creating hype about research and innovation. Instead, the process of policymaking in AI should be open to wider discussion and debate, as AI has serious societal implications. The Galgotias episode is a grim reminder.

Robot dog debacle raises tough questions

The hype and overemphasis on startups have resulted in our institutions making tall or fake claims, as Galgotias has done

Dinesh C. Sharma

Unlock Premium Insights in This Article