The Alarming Discovery
India's nodal cybersecurity agency, the Indian Computer Emergency Response Team (CERT-In), has issued a critical advisory for corporations across the country. According to its latest findings, there is mounting evidence that generative AI agents and tools
are a significant vector for corporate data leakage. The report highlights that employees, in their quest for efficiency, are feeding sensitive internal information into publicly available AI platforms. This information includes proprietary source code, confidential financial figures, unannounced marketing strategies, and even sensitive employee data. Unlike a traditional cyberattack involving a malicious hacker, this new threat is more subtle and stems from the very design of the helpful tools millions are now using daily.
How Productivity Tools Become Data Leaks
The root of the problem lies in how many large language models (LLMs) operate. When an employee uses a free or public AI tool to summarise a meeting, debug a piece of code, or draft a sensitive email, the data entered is often sent back to the AI provider's servers. This data can then be used to train future versions of the model. In effect, the confidential information is absorbed into the AI's vast knowledge base, becoming part of its operational data. The AI 'agent' is not acting maliciously; it is simply doing what it was built to do — learn from the inputs it receives. For the company, however, this means its trade secrets are no longer secret, posing an unprecedented and hard-to-trace risk.
A New Kind of Insider Threat
Cybersecurity experts are framing this as a new evolution of the 'insider threat'. Historically, this term referred to a disgruntled employee intentionally stealing data. Today, the threat comes from well-intentioned staff who are simply unaware of the security implications of their new workflow. They see a powerful tool that helps them do their job faster and better. They do not see the invisible data pipeline sending their inputs to a third-party server, potentially outside of India, with different data protection standards. This makes traditional security measures, which focus on preventing external breaches, less effective against this type of data exfiltration.
The Tangible Business Risks
The consequences of such leaks are severe and multifaceted. Competitors could potentially gain access to strategic plans or proprietary algorithms, eroding a company's market advantage. The exposure of financial data could impact stock performance or invite regulatory scrutiny. Furthermore, with India's Digital Personal Data Protection (DPDP) Act now in effect, leaking personally identifiable information (PII) of customers or employees, even unintentionally, can lead to substantial financial penalties and reputational damage. The CERT-In findings serve as a stark reminder that the convenience of AI cannot come at the expense of fundamental data governance and security.
CERT-In's Recommended Path Forward
Alongside the warning, CERT-In has outlined a series of urgent recommendations for Indian businesses. The primary guidance is the immediate development and enforcement of a clear corporate policy on the use of public AI tools. Companies are advised to either block access to non-approved AI platforms on corporate networks or invest in enterprise-grade AI solutions. These private AI environments can be sandboxed within a company's own secure infrastructure, ensuring that no data ever leaves their control. Crucially, the agency stresses the need for widespread employee education. Staff must be trained to understand what constitutes sensitive data and which tools are safe to use, turning them from an unintentional risk into the first line of defence.
