Artificial intelligence is changing the cybersecurity sector at an unprecedented pace. While AI is helping organizations strengthen threat detection and response capabilities, it is also giving cybercriminals new tools to launch sophisticated attacks at scale. As a result, cybersecurity has evolved beyond a purely technical concern and is now a critical business and trust issue.
In this episode of the Analytics Insight Podcast, host Priya Dialani speaks with Shrikrishna Dikshit, Partner, Risk Advisory at Baker Tilly ASA India. The discussion explores how AI is transforming cyber threats, why organizations must rethink traditional security models, and how continuous governance is becoming essential in an increasingly digital world.
Ans: I am with Baker Tilly ASAE DLLP, which is part of the global Baker Tilly network, one of the world's leading professional service organizations. The firm works closely with businesses across audit, tax, advisory, risk, and governance, helping them navigate complexity and make confident, strategic decisions.
Ans: I lead digital and cybersecurity practice, where my focus is on helping clients understand cyber risks, not just as technology issues, but also as business and trust issues, especially as the age of AI and in the age of AI, we work with an intention, intersection of cybersecurity, digital transformations, governance, and resilience, supporting clients on everything from cyber strategy and risk assessments, to incident readiness, regulations alignment, and AI era trust challenges.
Ans: AI has removed the skill barrier for attackers. Phishing today is now highly personalized. Voice, as well as identity, can be convincingly spoofed. Attacks operate at machine speed. That was not the case earlier. At the same time, defenders are using AI to detect behavioral anomalies and respond in real time.
The result is that cybersecurity is no longer people versus systems. It's an autonomous system on both sides, with humans setting the strategy and governance. So let me start, or let me give you a short story or a quick story.
Ans: AI removed these limits. Attackers can now generate thousands of highly personalized attacks simultaneously and adjust them in real time based on how it works. Hence, AI-driven cyberattacks are more dangerous than conventional ones. For one fundamental reason, they remove human limits from the attack. So, traditional cyberattacks relied on volume, technical skills, and time. AI collapses all three.
Over the past years, U.S. companies have reported a surge in cases where finance and operational staff are receiving phone calls from what sounded exactly like their CEOs, CFOs, same voice and mannerisms, same urgency. These are typically the wishing attacks, which we usually call them in our day-to-day scenarios.
Ans: AI introduces an entirely new attack surface that traditional security controls were never designed to protect. First, AI systems can be manipulated through prompt injections, data poisoning, or indirect inputs to behave in unintended and often harmful ways. Also, AI centralizes trust. When a model is connected to data, workflows, or payment systems, a single failure can amplify risk at machine scale.
Third, AI blurs the boundary between trusted and untrusted interactions, making it difficult for systems to distinguish legitimate commands from malicious ones. So to safeguard AI infrastructure, businesses must shift from static security models to continuous governance models.
To know more, listen to the full podcast.