Insurers need stronger safeguards against AI-driven fraud, say experts

Business Standard BFSI Insight Summit 2026: Insurers say AI-generated documents and evolving cyber threats are testing traditional risk assessment, prompting calls for stronger safeguards

Insurers need stronger safeguards, better data intelligence and human oversight to tackle emerging cyber risks as artificial intelligence (AI) makes fraud more sophisticated and traditional risk-assessment methods struggle to keep pace, industry experts said at the Business Standard BFSI Insight Summit 2026.

 

The panel discussion titled “Climate & Cyber Risks: Can Insurance Keep Pace With India’s Emerging Threats?” featured Animesh Das, managing director and chief executive officer of Acko General Insurance; Parthanil Ghosh, MD and CEO of HDFC Ergo; Pallavi Malani, India Leader and Insurance Partner; and Hitesh Joshi, chairman and managing director of GIC Re. The discussion was moderated by Subrata Panda of Business Standard.

 

AI complicates cyber-risk assessment

During the discussion, Animesh Das said that cyber risks differ from traditional insurance risks because a single underlying cause or technology provider can affect multiple retail and enterprise customers simultaneously. This makes it difficult for insurers to assess the potential scale of losses using conventional underwriting methods.

 

“Traditional point-in-time questionnaire-based underwriting cannot keep pace with risk changing almost daily,” he said.

 

He added that insurers would need to improve their ability to identify and prevent emerging risks rather than rely solely on historical data and existing assessment models. 

 

Hitesh Joshi highlighted the difficulty of assessing cyber-risk accumulation, as a single incident can affect organisations across geographical boundaries. He said preparedness varies among insurers and depends on their enterprise risk management systems, risk appetite, reinsurance arrangements and ability to stress-test potential losses.

 

The growing use of AI introduces additional uncertainty, he said, as insurers may face risks that are not adequately captured by historical data. This makes capital management and the assessment of potential losses particularly important.

Synthetic documents pose fraud risks

The panellists also discussed how AI and large language models (LLMs) could be used both to identify fraud and to create synthetic data and documents that make fraudulent activity harder to detect.

 

Das said new technologies and distribution channels have always required safeguards, adding that synthetic documents were not exclusive to the digital era.

 

“Whenever technology, whenever new lines of business, new channels come, we always create guardrails,” he said.

 

He stressed that insurers must continue to rely on data intelligence and sound underwriting practices to assess risks, even as the methods used by fraudsters evolve. Human oversight, he added, would be particularly important as insurers deploy more advanced AI systems.

 

“The human-in-the-loop concept is not only a statement, but it is actually a matter of fact,” Das said.

 

Ghosh said demand for cyber insurance was also expanding beyond technology companies, with large corporates and businesses in traditional industries increasingly discussing coverage. However, insurers need appropriate underwriting safeguards to ensure that businesses do not neglect basic security measures and subsequently rely on insurance to cover losses.

 

He said enterprise cyber insurance remains heavily dependent on reinsurance, while emerging risks associated with agentic AI require insurers to reassess existing products and coverage terms.

AI changes how customers discover insurance

The discussion also highlighted how AI is changing the way consumers search for and evaluate insurance products, creating new opportunities and challenges for insurers.

 

Das said the shift from search engines to LLM-based discovery was still evolving, requiring insurers to improve how they present and position their products across digital platforms. He said his company currently receives around 5-10 per cent of its leads through LLMs and related discovery channels.

 

Pallavi cited consumer research in which 45 per cent of respondents said they used LLMs for initial product discovery. However, consumers typically checked insurers’ websites after shortlisting products and then consulted agents, human representatives or call centres before making decisions.

 

“The pathways of consumer discovery are definitely changing,” Malani said, adding that insurers would need to develop more integrated, omnichannel customer journeys.

 

She said consumers were not simply relying on AI-generated responses, but were also verifying information through insurers and other channels.

Insurance gap remains among SMEs

The panellists also discussed the insurance gap among small and medium enterprises (SMEs), particularly in tier-two and tier-three towns, where products such as cyber insurance remain underpenetrated. Responding to a question on agent training, they said the industry needed to expand its reach to insurance distributors in smaller towns, noting that training had increasingly moved online.

 

On climate risks, the speakers called for better risk mitigation, affordable insurance products and stronger coordination between policyholders, insurers and reinsurers. The discussion also touched on urban flooding, drainage infrastructure and the need to improve the assessment of catastrophe exposures. The panellists also debated whether insurers should have greater flexibility in investing their funds to improve returns and potentially lower policy costs. While some acknowledged the benefits of investment flexibility, others cautioned that pursuing higher returns could expose insurers to additional risks.