IBM Research AI chief says India team drives our global AI ambitions

“Our research team here is actually at the heart of much of the coding effort globally. The Indian software team is also leading the development of the digital workforce automation tool watsonx Orchestrate and the data component of the IBM watsonx platform watsonx.data.” Vice President of Artificial Intelligence, IBM Research Sriram Raghavan said. Mint in an interview.

Raghavan, who leads a global team of more than 750 research scientists and engineers across all IBM research centers including India, was attending the company’s annual country-specific flagship event in Mumbai this year.

India is like a microcosm of IBM. Every part of IBM is here – research labs, software labs, systems labs, and we’re growing.

For example, IBM, a partner in India’s AI Mission and the country’s Semiconductor Mission, has installed WatsonX on the Center for Development of Advanced Computing (C-DAC) Airavat graphics processing unit (GPU) infrastructure for “startups and ecosystems” Partners are “available”.

On September 23, Indian Prime Minister Narendra Modi met with top technology leaders in New York, including IBM Chief Executive Officer (CEO) Arvind Krishna and Google CEO Sundar Pichai. He discussed with them topics such as artificial intelligence, quantum, biotechnology and life sciences, and semiconductor technology.

Also read: What turned IBM from tech giant to cautionary tale

Raghavan noted that IBM has a strong public-private partnership ecosystem in New York, with its lab in Albany working closely with the State University of New York and the New York State Nanotechnology Center. “We are learning from this and helping the Indian government build a similar ecosystem,” he explains.

Nearby, IBM is partnering with L&T Semiconductor Technologies Ltd to combine its expertise in semiconductor intellectual property (IP) with L&T’s industry knowledge to promote innovation in semiconductor solutions.

Artificial Intelligence Evolution

Raghavan emphasized that AI is gaining traction across hardware, programming and enterprise applications. “Companies need fit-for-purpose models that are efficient, scalable and affordable, and that’s IBM’s focus,” he said.

According to him, IBM’s artificial intelligence approach includes three key elements: Granite (IBM’s flagship brand of open and proprietary large language models (LLM)) series; the InstructLab open source project for customized models; the watsonx platform for cross-different environments Integrate, manage, and securely deploy AI models (including on-premises, public cloud, or IBM Cloud).

That said, like Meta Platforms Inc., IBM is committed to a public open source model. “The real value is in managing and optimizing these models, just like we do with Red Hat and Linux,” Raghavan said.

When Gen AI emerged, it raised concerns that models were closed, proprietary, and dangerous.

“So, we (IBM and Meta) launched the Artificial Intelligence Alliance (December 2023) to highlight the value of an open approach, and many Indian companies have joined this movement, recognizing that AI is too important to be developed behind closed doors ,” Raghavan said.

Also read: Quantum-centric supercomputers are about to become a reality: IBM’s Dario Gil

The Artificial Intelligence Alliance currently includes IIT-Bombay, AI4Bharat (IIT-Madras), IIT-Jodhpur, Infosys Ltd, KissanAI, People+AI and Sarvam AI.

“By keeping the model open, we draw more eyes to help innovate and build better safeguards. It’s not the model that creates risk, it’s how it’s used,” insists Raghavan.

He emphasized that the U.S. government recognized this approach in its recent executive order and understood that overly restrictive measures can stifle innovation, especially in academia and startups.

Raghavan believes the underlying technology should be open to facilitate collaboration and drive new ideas, even though customers will still pay for enterprise-grade support, security and management. “Monetization will come from managing AI applications,” he explained.

But will enough companies move from pilot to production, and how will they get return on investment (ROI) from GenAI? “As customers move from proof of concept (POC) to production, our top priorities are cost, performance, security and skills,” asserts Raghavan. He cited an IBM study that showed 10-20% of companies have expanded at least one AI use case. He acknowledged the numbers were growing but challenges remained, particularly in regulated industries.

Also read: Let’s see if AI can work its magic to make up for the educational shortcomings

“Successful companies focus on key areas with clear return on investment potential rather than spreading work across multiple POCs. This targeted approach allows them to scale efficiently and achieve meaningful returns. As companies scale their workforce Intelligent use cases, they see the importance of balancing technology, process and culture,” he said.

Artificial Intelligence Use Cases

According to him, IBM sees three key use case categories: customer service, application modernization, and digital workforce and business automation. “Even before Gen AI, customer service was a natural choice. Everyone wanted better customer service at a lower cost. The real value comes from creating fit-for-purpose models that fit specific needs,” he explains Dow, adding that, for example, a customer service model does not require complex problem solving, which helps reduce costs.

Application modernization is also critical, especially when enterprises are dealing with large legacy code bases. “For example, IBM’s Watsonx code assistant for COBOL, an ancient programming language, helps modernize mainframe code and makes it easier for developers to work with older languages. We are extending it to Java, which is another A language that is critical to businesses. Digital workforce or business automation covers processes such as supply chain, finance and human resources. Our Watsonx Orchestrate suite is designed to streamline these operations using artificial intelligence,” he elaborates.

However, Raghavan acknowledged that as businesses adopt AI, they face challenges in three areas: skills, trust and cost. “This is where watsonx.governance comes in – it helps automate model governance, ensuring correct usage, tracking data and running risk assessments.”

When asked about the debate surrounding AI’s enhanced “reasoning” capabilities, Raghavan admitted it was “a delicate subject.” Models don’t reason with logic like humans do, he explained. Instead, they learn by example. “While current AI can reason in specific domains such as IT systems or code, general reasoning remains out of reach.”

Also read: GenAI has a killer app. This is coding, says Naveen Rao, Head of Databricks AI

Domain-specific reasoning is also “very useful,” Raghavan said. For example, AI can improve IT automation or help solve coding problems by learning from examples, making it a practical and valuable approach.

He concludes: “We are also seeing a shift from models that simply provide answers to those that ‘think’ before reacting. These models that engage in System 2 behavior (like Daniel Kahneman’s analogy) can be self-critical and refine their responses. This will drive more complex AI tasks, but at an increased cost because inference time increases with deeper inference.”