Written by Aditya Kalra and Munsif Vengattil
NEW DELHI (Reuters) – India will ask tech companies to seek approval for “untrusted” or under-trial artificial intelligence (AI) tools before making them available to the public, urging users to He also pointed out that the possibility of giving an incorrect answer should also be labeled. query.
The use of such tools, including generative AI, and “making them available to users on the Indian internet must be done with the express permission of the Indian government,” the country’s IT ministry said. , said in an advisory issued to platforms last Friday.
Countries around the world are competing to create rules to regulate AI. India has tightened regulations on social media companies, which consider the South Asian country a top growth market.
The recommendation comes in response to Indian Prime Minister Narendra Modi, who has been accused by some of implementing policies characterized as “fascist”, after government leaders announced on February 23 that Google’s AI tool The announcement comes a week after he slammed “Gemini.”
The next day, Google moved quickly to address the issue, but said the tool was “not always reliable”, especially when it comes to current events and political topics.
“Security and reliability are legal obligations of platforms. ‘Sorry, we can’t trust you’ is not an exemption from the law,” Rajeev Chandrasekhar, deputy IT minister, said in response to Google’s statement on social media platforms. X mentioned above.
India’s Friday advisory also called on platforms to ensure their AI tools “do not threaten the integrity of the election process.” India will hold a general election this summer, with the ruling Hindu nationalist party expected to secure a majority.
(Reporting by Aditya Kalra in New Delhi and Munsif Vengattil in Bengaluru; Editing by Christian Schmollinger)