AI is changing how financial institutions detect fraud, manage operations and serve customers, but it is also giving cybercriminals faster and more precise ways to attack. Yogesh Jadhav, CTO at Choice International, discusses the security risks of AI, the need for stronger governance and how financial institutions can prepare for evolving threats.
In today’s episode of the Analytics Insight podcast, host Priya Dayalani speaks with Yogesh Jadhav, CTO at Choice International, about the growing role of AI in financial services and the cybersecurity risks that come with it.
Yogesh explains his point of view about how generative AI is making the cyberattacks happen quicker, smarter, and harder to detect. He talks about how difficult it is to secure nondeterministic AI, the necessity of access control, audit trails, human oversight and the need for cybersecurity to be considered during the entire lifecycle of the technology.
Some other topics that have been covered in this interview include AI-based financial scams, phishing attacks, security exercises and real-time governance. Yogesh gives insight on how India's digital payments environment can help give rise to the next generation of AI and cybersecurity innovations. Here are some important excerpts from the interview.
Generative AI has made cyberattacks faster and more convincing. Attackers can personalise phishing messages, use local languages and target people more precisely, making it harder for financial institutions and customers to spot scams.
The performance of AI is also unpredictable at times, in the sense that the same input will yield varied results, thereby making it difficult to test and monitor.
Yogesh suggests that organizations should look at their AI system like a new hire who should be given an identity, access should be limited, it should be monitored, and humans should play an active role in decision making through reviews and maker-checker systems.
Organizations can conduct realistic phishing and attack simulations to know what AI-powered attacks will look like. Besides that, AI can be used to help security teams filter out irrelevant information and concentrate on real threats.
India's massive digital payments ecosystem gives organisations experience at a scale few countries have matched. The data, transaction volumes and fraud-prevention systems being built here could help shape the next phase of AI security globally.