BENGALURU: Though artificial intelligence (AI) can perform tasks autonomously, there should be mechanisms for human intervention, especially in critical situations or when ethical dilemmas arise, says Rohan Malhotra, CEO and founder, Roadzen. In an interaction with TNIE, Malhotra says there is a need to identify the grey areas where a direct true/false decision can’t and shouldn’t be taken by AI. “Identifying cases which belong in this grey area can be the first step towards a solution,” he says.
The company utilises AI/ML to empower auto insurance life cycles for a long time as it is a tech-driven company. “Any Machine Learning solution requires good quality data to deliver its impact. Often, the dataset required to solve certain use-cases is not adequate. Roadzen leverages Generative AI to bridge the gap in data. We are able to model after and generate synthetic datasets for object classes not found in abundance while processing claims,” he explains.For instance, rust on vehicle parts is often scarce in the dataset, while examples of scratches, dents and broken parts are in abundance. This could lead to the class-imbalance problem in AI. He adds that Gen AI helps them mitigate the same.
Rohan Malhotra
CEO and Founder, Roadzen