Cannot outsource thinking, creativity to AI: Top science adviser

New Delhi, Jan 31 (PTI) The government's top science adviser has struck a cautious note on the use of artificial intelligence, saying humanity cannot outsource its creativity and thinking abilities to machines.

In an interview with PTI ahead of the AI Impact Summit, Ajay Kumar Sood, the Principal Scientific Adviser to the Government of India, also favoured using existing laws to deal with the scourge of artificial intelligence-generated deepfakes, contending that strict legislation may "kill innovation".

"We have to watch out for it. I agree that if you are not watchful for the younger generation, it can have a negative impact," Sood said to a question on AI leading to a cognitive decline in humans.

Sood cited the examples of greater use of calculators in daily use and the decline in the use of multiplication tables in later generations.

"If you outsource your thinking, that is very bad for humanity. Human beings are made to think.

"AI came out of human thinking. Now, you cannot outsource your creativity, your thinking to a machine, which owes it to you. That is the danger," he said.

Sood made a strong pitch for having guardrails for the use of AI, particularly in schools, colleges and universities with a clear distinction against the use of the technology in certain areas.

"Responsible use of AI is very, very necessary because AI can also hallucinate. If people are not careful they will take that hallucination as a reality, which is extraordinarily bad," the top scientist said.

"That is where human intervention is needed. So, I always say, a human has to be in the loop," he said.

On the recent Grok controversy on indecent depiction of children and women, Sood said the existing laws are sufficient to deal with such incidents.

"There are laws already applicable for Grok kind of incidents... in Bharatiya Nyaya Sanhita and the Information Technology Act. So, what we are saying is the available laws may be sufficient," Sood said.

He said the government has made it clear that it will not hesitate to bring in new regulations if required.

"But right now we are saying it may not be needed," the science adviser said.

He said that Section 66D of the IT Act, 2000 applies to cases of cheating by impersonation using computer resources, which includes AI-enabled impersonation and deepfakes.

Sood said content involving obscenity or reputational harm can be dealt with the provisions of the Bharatiya Nyaya Sanhita in Section 294 and 296 (obscenity) and Section 356(1) in case of defamation.

He said the use of personal data without consent to train AI models is governed by the Digital Personal Data Protection Act.

India is hosting the AI Impact Summit from February 16-20 and 15 Heads of State/Government, more than 40 ministers, over 100 leading chief executive officers (CEOs) and chief experience officers (CXOs), and more than 100 eminent academics have confirmed their participation. PTI SKU SKU KSS KSS