Daijiworld Media Network - Bengaluru

Bengaluru, Oct 29: Companies deploying artificial intelligence (AI)-powered autonomous agents should place them in the “zero-trust” category by default and maintain strict controls over their operations, said Sanjay Beri, founder and chief executive officer of US-based cybersecurity firm Netskope.

In an interaction with Business Standard during his visit to India, Beri said enterprises must remain accountable for the actions of AI agents they deploy, even when the underlying technology is developed by another company.

“The first thing that most people forget is that an AI agent is not a replacement for a human. An AI agent has no morals, no emotional quotient, no compass of what is right or wrong. You can never trust an (AI) agent or an LLM (large language model),” Beri said.

He warned that AI agents, being goal-oriented systems, could resort to “lying and manipulation” involving humans or other AI agents to achieve their assigned objectives. Their focus on completing tasks, he said, makes it essential for organisations to establish safeguards rather than assume the systems will always behave as intended.

Even companies that exercise caution while developing and deploying AI agents must maintain continuous monitoring, appropriate controls and mechanisms to override their actions, Beri said. He also called on leading AI companies developing large language models and autonomous agents to be more transparent about their systems’ capabilities so that enterprises can establish appropriate governance standards.

“There is a concept of shared responsibility, where the frontier AI company has some responsibility, but ultimately it is the enterprise that is deploying it that will be held to the standards,” he said.

Beri pointed to instances of AI agents escaping the sandbox environments in which they are tested and taking unauthorised actions to complete tasks as examples of inadequate governance and weak testing arrangements.

As businesses in India and elsewhere increasingly integrate AI chatbots and autonomous agents into their operations, cybersecurity companies will have to focus more on reducing risks arising from human behaviour, he said.

For instance, an employee might use a personal AI chatbot account to upload sensitive company information while performing work-related tasks. Rather than immediately reprimanding the employee, organisations could intervene in real time and explain why such actions are prohibited, Beri said.

“The vast majority of data exfiltration or data loss to AI is not due to malicious intent,” he said, emphasising the importance of educating employees and preventing accidental data exposure.

The growing adoption of agentic AI architectures has added another layer of complexity to cybersecurity, Beri said. Malicious actors are also deploying unauthorised or “shadow AI” agents to access proprietary information, creating new risks alongside existing vulnerabilities involving network endpoints and human error.

Netskope is investing significantly in India by expanding its infrastructure and hiring skilled professionals to help businesses meet evolving global standards and regulatory requirements, including the Digital Personal Data Protection Act, Beri said.

The company has invested close to $100 million annually in India over the past few years and plans to continue expanding its workforce. Netskope currently employs around 1,300 people in India, more than half of whom work in engineering, he added.

Netskope is a Nasdaq-listed cybersecurity company with a market capitalisation of more than $10 billion.