The world has shown excitement – and skepticism – about the rise of GenAI and large language models (LLMs). Jim Goodnight, a leading figure in the field of AI and data analytics, believes GenAI is powerful but cannot solve business risks on its own. Goodnight, CEO of North Carolina-based SAS, which he co-founded in 1976, tells Sudhir Chowdhary in an interview in Orlando what GenAI means for business analytics, the buzz around LLMs and how India can advance its own AI infrastructure. Excerpts:
As a technologist, what excites you the most? What is the next big technology you are looking at?
It’s a tough choice, given how rapidly technology is evolving. But two areas stand out: Generative AI and quantum computing. At SAS, we’re expanding our trustworthy GenAI capabilities – especially in synthetic data generation, LLM governance, and orchestration using SAS Viya and our Viya Copilots. At the same time, quantum computing is fast emerging as the next major technological frontier. Our R&D teams are actively collaborating with leading partners and some large customers to explore how quantum computing can address complex problems across industries.
How do you see the current hype around LLMs like ChatGPT, DeepSeek and Gemini?
The current hype is driven by a shift in focus from traditional, rules-based AI to generative AI, which brings language-based models into the spotlight. At SAS, we work with both large and small language models – some with just a few billion parameters – to enable use cases like generating SAS code from natural language, explanations of bias in AI models or embedding LLMs into automated decisioning systems.
For example, in a healthcare setting it would be used to summarise patient records, automate clinical documentation or enhance patient engagement. In banking, LLM’s are used to parse, summarise and extract risk indicators from unstructured documents like credit memos, regulatory filings, or contracts. Organisations are investing heavily to leverage GenAI, but our advice remains: stay open-minded, flexible, and cautiously curious. Data privacy and security should remain a priority – especially in regulated sectors. That’s been our guiding principle as we help customers operationalise GenAI responsibly.
DeepSeek has intensified US-China race for supremacy? How do you perceive this technology conflict between the two countries?
The rise of platforms like DeepSeek is expected. In fact, our 2024 global GenAI study highlighted that China leads the world in GenAI use overall, and the United States leads the world in full GenAI implementations. We are particularly curious about what value, and risks, it brings for those customers in highly-regulated industries like banking, financial services, insurance, government and healthcare where data privacy and security are top concerns. While LLMs can be built using responses from existing models, what matters more is the rigor in governance, transparency, and alignment with ethical frameworks.
Should India build its own LLM? How can it create the ecosystem for tech companies to thrive?
Yes, India should invest in building its own LLM. The barriers to entry have lowered significantly- there’s a wealth of open-source models, datasets, and documentation available. But more importantly, India has unique linguistic diversity, data needs and industry contexts that global models may not fully address. To succeed, we need a collaborative ecosystem with strong public-private partnerships, investment in high-quality data infrastructure, localised datasets, academic research support, and clear regulatory frameworks. Encouraging innovation while ensuring ethical, transparent, and inclusive development will be key in the GenAI era.
How do you see AI progressing in the next two to three years?
We’ve already seen tremendous advances since late 2022, especially with generative models becoming more grounded and context-aware. While general intelligence remains a distant goal, today’s models are getting better at producing relevant outputs when prompted correctly. Commercially, we see huge potential in areas like fraud detection, customer analytics, predictive maintenance – and yes, even drug discovery. AI can potentially help identify promising compounds, speed up clinical trial design and improve patient outcomes. But broader adoption depends on demystifying AI for the user. People need clear, factual knowledge about AI – what it is, how it works, and where it truly adds value.
How do you think about the issue of AI and jobs?
Like any transformative technology, AI will displace some roles – but it will also create many more. But it can boost overall productivity for most workers and propel economic growth in the long run. AI and advanced analytics lead to better-run organisations that are more profitable and more competitive. This leads not only to bottom-line growth, but to more opportunities for workers. AI will also spark new jobs and new careers. In manufacturing today, for example, new AI- and technology-infused jobs range from data scientists and process engineers to precision assemblers and robotics technicians.
(The correspondent was in Orlando at the invitation of SAS)
