I will never allow AI to take over my thinking or writing: Sridhar Vembu

Zoho unveiled its biggest AI push yet at its Zoholics India event in Chennai on Wednesday (September 23), launching Zia Chat, a conversational AI interface designed to work across Zoho applications and connect with external business systems.

Speaking to The Federal on the sidelines of the event, Zoho founder Sridhar Vembu discussed how he uses AI and how the company expects its engineers to work with the technology.

‘I do not allow AI to take over thinking'

Vembu said he is a heavy user of AI but does not allow it to take over the thinking or writing process. He said he writes his own social media posts, presentations, and internal communications, while using AI to review, research, and provide information.

“I take full responsibility for it,” Vembu said, adding that he uses AI to assist him but does not ask it to write his content.

Also Read: AI doom fears overstated, global pause unrealistic: Zoho's Sridhar Vembu

He compared this approach with how software engineers should use AI to generate code. According to Vembu, engineers need to understand the overall system, objectives, and technical approach, and must be able to guide the AI rather than simply accept what it produces.

‘Human is responsible’

Vembu said Zoho's approach is to ensure that programmers remain responsible for every software release, regardless of how much AI is used during development. “The AI is not responsible. You, the human, are responsible,” he said.

Also Read: India may beat America in AI adoption: Sridhar Vembu

He described AI as a tool that can help employees perform their jobs better, but stressed that the technology should not be allowed to take over the job itself.

For Vembu, the key issue is retaining the underlying thought process and understanding, while using AI to improve productivity.

‘Reckless adoption is a risk, not the technology itself’

Vembu was asked about the recent departures of AI researchers from companies including Anthropic and Google DeepMind, and the concerns raised about AI becoming dangerous.

Also Read: ‘Globalisation is dead’: Vembu, Pai call for stronger AI push after Anthropic move

He said he sees the greater risk in the reckless adoption of AI in areas where it may not be appropriate, rather than in the technology itself.

“Reckless adoption is a risk, not the technology itself,” Vembu said.

Need for safeguards

Vembu compared AI with any other powerful technology, arguing that the risks increase when such technologies are deployed without adequate safeguards.

He said AI companies need to debate the safeguards around their technology internally, while governments and regulators can establish appropriate liabilities for its use.

He gave the example of driving: if a person crashes a car, the person driving is held responsible. Similarly, he argued that if someone uses an AI system and that AI causes harm, there should be a clear framework establishing responsibility.

Also Read: Zoho’s Sridhar Vembu warns coders to seek alternative careers amid AI boom

“If I'm using AI and the AI hurts somebody, I have to be responsible,” Vembu said.

He said such laws and liability frameworks could help establish clearer accountability as AI becomes increasingly integrated into products and businesses.

Zoho's latest AI push therefore comes with a message from its founder that the technology can be adopted aggressively, but responsibility for how it is used must remain with humans.