Has India Chained the AI Revolution

A new regulation announced by India that requires technology companies to seek prior approval before public launch of AI-led tools that are under trial appears to have put it on par with similar steps taken by China and the European Union. The rules, which require tech companies to not permit bias or threaten the country’s electoral process, makes the US approach seem feeble.

The advisory, which came from the Ministry of Electronics and IT, is not binding, says junior minister Rajeev Chandrasekhar, but does signal that this is the future of the regulation. “We are doing it as an advisory today asking you to comply with it,” he said in a post on X. He sought to mollify the industry by suggesting that it was aimed at “untested AI platforms deploying on the India Internet” and that it doesn’t apply to startups. 

What has prompted India’s AI reversal 

By the looks of it, the sudden about turn from the government has come about due to two factors and another that is directly associated. The first of these relates to Google’s Gemini fiasco where the GenAI tool was found spewing inappropriate content, including a biased response against Prime Minister Narendra Modi, which the company later apologized for. 

The second obvious link is the upcoming general elections where voters in India would select 543 lawmakers across the country in what is seen by the ruling party as a referendum of sorts for its economic policies over the past decade. The advisory specifically refers to this point stating that services or products must not “permit any bias or discrimination or threaten the integrity of the electoral process.” 

The third associated factor could be Google’s recent announcement about a tie-up with a consortium of news publishers and fact-checkers to curb misinformation. In a blog post, Google said it would support these publishers to combat fake news and create a common repository that news publishers can then use to tackle the misinformation challenge. 

The swings and roundabouts of the advisory

Though the minister said that the advisory was not binding, he also notes that the power given to the government via the IT Act 2000 and the IT Rules, 2021 would apply in this regard. It also seeks compliance with “immediate effect” while also seeking action-taken reports and status reports within 15 days. 

The advisory also puts the onus on the “significant” tech firms to “appropriately label the possible and inherent fallibility or unreliability” of the output that AI their AI models generate. To the discerning, this marks a clear reversal of India’s earlier hands-off approach to AI regulation where the government refused to regulate AI growth while identifying it as a strategic one for India’s continued global contribution to the IT industry.

As recently as in December, India’s IT minister Ashwini Vaishnaw had told delegates at a Global Partnership on Artificial Intelligence Summit that AI could potentially address potential issues and improve lives across the world. “AI also presents challenges as it is not amenable to geographical boundaries and disruptive technologies raise the question of how society should respond to such development,” he had said while reiterating India’s commitment to democratizing technology and its use for inclusive growth. 

How is the industry reacting to this announcement?

Many of them think the move isn’t kosher and admit that it has taken them by surprise. Many have taken to X (formerly Twitter) to share their views. “I was such a fool thinking I will work on bringing GenAI to Indian agriculture from SF. We were training a multimodal low cost pest and disease model, and so excited about it. This is terrible and demotivating after working 4 years full time bringing AI to this domain in India.” he said. 

Others like Bindu Reddy, CEO of Abacus AI and former employee of Amazon and Google, have been forthright claiming “India just kissed its future goodbye!” Her post on X says, If you know the Indian government, you know this will be a huge drag!  All forms will need to be completed in triplicate and there will be a dozen hoops to jump through! This is how monopolies thrive, countries decay and consumers suffer! Sadly India is already dominated by monopolies, nepotism and bureaucracy and this new rule just made it far worse.” 

Eminent digital rights activist Nikhil Pahwa has questioned the veracity of the advisory asking the following questions via X: (a) how an advisory is legally binding. Not an ordinance; (b) how it is backed by the IT Act plus legally kosher © how it excludes startups and (f) what does untested mean. He also re-shared another post that listed 533,534 open source AI models that would require approvals. 

While it appears that the government seems to have created a flutter over what was entirely the fault of Google’s chatbot, by the looks of things the law could well become a bit more benign once the risk of deep fakes messing up with India’s general elections recede. Therefore, it may be a good idea to wait and watch till after May 2023.