The Double-Edged Sword of AI in Cyber Insurance
The integration of Artificial Intelligence (AI) into the insurance industry is not a novel concept, particularly in the realm of cybersecurity. However, as cyber threats and ransomware attacks evolve at a staggering pace, organizations are compelled to continually enhance their strategies, not just to maintain a competitive edge, but also to ensure their very survival.
A recent Roundup Report 2024 by Cowbell highlights that larger enterprises — specifically those with annual revenues exceeding $50 million — are at a higher risk, facing cyber-related incidents at a rate 2.5 times greater than their smaller counterparts. In a conversation with Insurance Business, Rajeev Gupta, the co-founder and chief product officer at Cowbell, illustrated AI as a double-edged sword. While it poses significant challenges, it also serves as an indispensable tool that organizations must adopt.
“Employing AI to identify fraudulent claims isn’t merely advantageous, it’s essential. Cybercriminals are already utilizing AI,” Gupta explained. “Claims are being submitted with AI-generated documents. An individual can craft a seemingly legitimate ransomware attack with all the markers being AI-created. If a human is tasked with evaluating that claim, they might be deceived into believing it’s authentic. This is where AI becomes crucial. Insurance firms that fail to leverage AI could find themselves on the losing end of this battle.”
The Battle of AI Versus AI
In this modern landscape, it’s a matter of AI pitted against AI. Cybercriminals are rapidly enhancing their skills, using sophisticated tools to circumvent fraud detection mechanisms. While it’s imperative for insurers to deploy AI, Gupta underscores the necessity of adhering to ethical boundaries.
“There are numerous constraints regarding what AI can achieve,” he remarked. “The complexity of certain risks makes it challenging to rely solely on AI. Human intervention remains crucial, particularly in intricate risk assessments, underwriting, and claims management.”
Gupta compares this human-AI collaboration to a Venn diagram, where both elements overlap to complement each other. AI excels at revealing patterns and insights that may elude human perception, thanks to its robust data processing capabilities.
“AI can synthesize information in a manner reminiscent of a professional with four decades of experience,” he noted. “Yet, it can achieve such expertise within mere months. However, certain tasks necessitate human intuition. A hybrid approach is thus optimal, blending the strengths of both AI and human intelligence to provide superior services to clients.”
Supporting this perspective is data from Duck Creek Technologies, indicating that 44% of consumers now prefer human interaction after their policy is established, a rise from 35% in 2022.
The Future of AI in Cybersecurity
“AI will undoubtedly improve with time,” Gupta added. “If 50% of underwriting is currently managed by AI systems, this figure could increase to 60 or 70%. Nonetheless, humans will remain indispensable for complex decision-making processes and strategic oversight. While the future remains uncertain, the necessity of human oversight to address bias, outliers, and uphold ethical standards will persist.”
Looking forward, Gupta emphasized the critical nature of maintaining accuracy and fairness in AI assessments within cybersecurity. Static models are insufficient in this dynamic threat landscape.
“Continuous assessment is paramount,” he stated. “Concerns about bias in AI arise — while initial data may be limited, constant evaluation and the integration of diverse signals into models help mitigate bias. This is crucial for maintaining model relevance. As threat actors are already exploiting these technologies, we require adaptive models capable of identifying new threat types. Continuous threat intelligence and business assessments are therefore vital.”
News
