Trust is essential to adoption of AI: Gargi Dasgupta of IBM Research India

In a Q&A, the firm's director dwells on the impact of climate change on business output and how her company's Environmental Intelligence Suite aims to reduce this risk

With climate change threatening to impact business output, too, IBM Research recently launched its Environmental Intelligence Suite, which has a significant contribution from its India team. Gargi Dasgupta, Director, IBM Research India, spoke to Shivani Shinde about the company’s use of artificial intelligence and analytics to reduce this risk, dwelling on the significance of the India team in IBM Research’s global roadmap. Edited excerpts:

How is the Environmental Intelligence Suite application different from what is already available in the market?

We are the first company to bring together AI, weather, climate risk analytics, carbon accounting and operational data together in a single SaaS solution to address climate risk and sustainability objectives.

We’ve used our expertise in AI, automation innovations and cloud, alongside data from The Weather Company (ranked as the world’s most accurate forecaster overall), and tied it with the technologies that clients are already using -- including asset management and supply chain technologies -- to provide them with valuable environmental insights.

To build this solution, IBM Researchers focused on two critical needs -- first, accurate analysis of climate risk at scale, and second, precise accounting of carbon emissions. To address the first, they built a new model that collects, analyses and maintains massive amounts of heterogeneous and unstructured data, and enables the prediction of the risk and potential impact of upcoming climate and weather hazards in a much more efficient, standardised and integrated way than current methods.

To address the second need, IBM scientists have designed new AI-enhanced, general-purpose carbon footprint-reporting, tracking, and optimisation capabilities that measure carbon emissions more accurately and automatically across the full spectrum of a business’s operations. These visualisations help teams to interpret source data, understand how these emissions were calculated and tracked, and where they can act to reduce emissions.

What is the contribution of the Indian research team in building this suite? Can you quantify this in any way?

Researchers across the globe work collaboratively to build innovative industry solutions, leveraging their expertise and experience. Last year, IBM research launched a global initiative called “Future of Climate”, to develop and demonstrate innovations focused on addressing climate change, leveraging its AI expertise. This EIS solution is one of the outcomes of that effort, which takes advantage of innovations built by IBM Research, including researchers from India.

 How are companies in India looking at climate-change related issues and what level of readiness are they at?

According to the CDP India report, Indian companies put the total inherent financial impact of climate risks at Rs 7,138 billion, with an average risk per company of Rs 92 billion. Up to 94 per cent of reporting companies have identified climate-related opportunities as having the potential to make a substantive financial or strategic impact on businesses.

We see these impacting organisations, leading to an increase in weather-related disruptions and costs. Organisations need actionable environmental insights to mitigate these risks and get ahead of these challenges. Today, many organisations across industries are capitalising on weather data to mitigate challenges that can impact their businesses.

How has IBM Research in India evolved over the years? What are its focus areas?

IBM Research India was one of the first industrial computer science labs, set up in India in 1998. Over the years we have built our expertise in the areas of AI and hybrid cloud, and are focused on finding solutions to grass roots-level challenges for India, with a potential to scale up at the global level.

Our AI research is focused on three areas -- natural language processing (NLP), automation, and trust. Through our NLP work, we help AI understand the language of business by advancing its capability to generalise, reason, perform sentiment analysis, and extract meaning from a variety of complex multi-format documents. These capabilities are then built into Watson offerings. We are also enabling our NLP solutions to understand different languages (14 so far, including Hindi, and other Indian languages in our road map are Bengali, Tamil, Telugu, Kannada, Marathi, Gujarati and Punjabi), such that they can extract useful information for conducting business.

We are using AI to transform IT and help our clients embark on their cloud journey. In hybrid cloud research we are focused on driving innovations that make our clients’ move to the cloud seamless, scalable, and secure. Our research work is also made accessible to the open-source community for wider adoption and acceleration of innovation. We are also applying our expertise to find more efficient, intelligent, and sustainable solutions to climate change-related risks.

Though AI is the buzzword, it also creates biases in the system. How do you make sure this does not happen in the system you are creating?

Early this year, the global research firm Morning Consult conducted a global AI adoption study on behalf of IBM. They found that 95 per cent of IT professionals in India believe it is critical or very important to their business that they be able to trust AI’s output as fair, safe and reliable.

Clearly, trust is essential to AI adoption. At IBM, we follow high-level principles for trust and transparency, to strengthen trust in technology, including AI. The principles make clear that the purpose of AI is to augment human intelligence; data and insights generated from data belong to their creator; and powerful new technologies like AI must be transparent, explainable, and free of harmful and inappropriate bias in order for society to trust them.

Organisations today recognise that it takes a holistic approach to manage and govern their AI solutions across the full AI lifecycle. We’ve contributed to the development of the open-source toolkit AI Fairness 360, which allows developers to detect and de-bias the AI models. AI explainability 360 provides technology that enables not just developers, but other persons like risk officers and end-customers to open up the AI model and understand the reasoning behind the decision. We have also introduced AI factsheets, which provide information about a model’s important characteristics, just like nutrition labels for foods or information sheets for appliances.

Supply-side constraints are impacting tech and IT services in India. Have they impacted IBM Research in India?

The presence of a large number of engineering institutes provides us with access to some of the best skills and talent. We drive various programmes and initiatives to nurture our talent pipeline. The requirements for a research lab are evolving and some of the skills are unique to the research domain. As a lab, we work closely with academia -- both students and professors -- on joint research projects, mentorship, internship programmes, and providing platforms for students to work on real-world problems. For example, our collaboration with IIT Bombay and IIT Delhi is part of IBM Research’s AI Horizon network -- a global initiative for joint research in AI -- to address real-life enterprise challenges or challenges that are unique to India.

In June this year, we collaborated with IISC Bangalore to launch the IBM IISc Hybrid Cloud Research lab. Our commitment is to grow a quantum-ready workforce and build an ecosystem to nurture the quantum community in India. And we’ve collaborated with top-tier institutions in India to accelerate advanced training and research in quantum computing.

How significant is IBM India to the global research team?

IBM Research is one of the world’s largest and most influential corporate research labs, spread over 17 locations across the globe. Since our first lab opened in 1945, we’ve authored more than 110,000 research publications. Our researchers have won six Nobel Prizes and six Turing Awards, and IBM has been granted more than 150,000 patents. Our mission is to ensure that IBM stays in the vanguard of technological innovation that helps enterprises achieve digital transformations.

Patenting is a collaborative effort, and hence we do not call out the numbers according to business units. However, we are proud to share that at the country level, with over 930 patents granted in 2020, IBM India has been the second highest contributor to IBM’s patent leadership (after the US) for four years in a row. And these patents are in emerging technologies like AI, cloud, blockchain and security.

How does the split of IBM impact IBM Research?

Our strong commitment to R&D and focus on fostering a steady pipeline of emerging innovation does not change with the separation of Kyndryl. IBM Research always has been, and will continue to be, a significant contributor to IBM’s broader commercial offerings available to clients.