Sridhar Vembu Warns Companies Against Sharing Secrets With AI Labs
Zoho chief scientist Sridhar Vembu has urged businesses to protect proprietary information and consider building their own AI models instead of sharing sensitive data with frontier AI labs.
Zoho chief scientist Sridhar Vembu has urged companies to be cautious about sharing sensitive business information with frontier AI labs. He argues that critical know-how, intellectual property, trade secrets and proprietary processes should remain under a company’s direct control, and suggested that businesses consider building their own AI models.
Vembu said companies should treat their internal expertise as something that belongs within their own AI systems. He also noted that the cost of training and operating AI models is declining, potentially making private AI infrastructure more accessible to businesses.
He cited concerns raised by mathematician Terence Tao following claims that an OpenAI model had solved the long-standing Navier-Stokes problem. Tao warned that increasingly capable AI could change incentives around sharing research, potentially discouraging scientists from openly discussing promising ideas.
Vembu highlighted the pharmaceutical sector as an example of companies using private AI systems, often built with open-source models, to keep sensitive information about drug development away from external AI platforms. He suggested that similar precautions could apply to companies whose competitive advantage depends on proprietary formulas, software or business processes.
The debate has also intensified after mathematician Tristan Buckmaster alleged that AI-generated work may have benefited from research ideas that were available online. While Vembu’s focus is corporate security and Tao’s concerns centre on scientific openness, both point to a broader issue: increasingly capable AI systems can rapidly process and exploit information once they gain access to it.
