Working to assess safety of AI; job loss not intense in India: IT Secy

The IT ministry is working on a mechanism to assess safety and trust elements in artificial intelligence (AI) solutions but does not want to start with a regulation as it may hurt innovation in the AI space, a top official said on Thursday.

Speaking at CII's Global Economic Policy Forum, Ministry of Electronics and IT Secretary S Krishnan said the government has broadly addressed key regulations barring a couple of the issues like copyright and existential crisis.

He said under the IndiaAI mission, the government has tried to make a subtle distinction between responsible AI as defined by the West -- which is not expected to be regulated by the government -- and safe and trusted AI.

"Safe and trusted AI is when you do the innovation, try and be responsible about it but there would be an assessment. We will try and see that it is safe and trusted in a manner that works for everybody. That is the other space that we are working on right now. We don't want to start with regulation, and regulation, in a sense, would actually hurt innovation in this particular space," Krishnan said.

He said there is a need for regulation around harms from AI, misrepresentation and deepfakes in the AI domain and there are enough laws in the country to act on misrepresentation.

Talking about the impact of jobs in India, Krishnan said the impact in the country is not as intense as in the west.

"We looked at the job loss element separately in terms of skilling people to take up jobs which could possibly supplant some of the other jobs. Our concerns with regard to job loss in India, particularly with generative AI, are probably not as intense as in the West, where office jobs and white collar jobs are a lot more than what they are here," Krishnan said.

He said there are concerns around the use of personal data that have been addressed in the Digital Personal Data Protection Act and the rules will be out soon.

Regulation around copyrights in India will be based on the courts' orders, Krishnan said, adding that there was no need to regulate AI for existential crisis.