Minister of State for Electronics and Information Technology Rajeev Chandrasekhar said on Monday that the government’s advisory on large language models will not apply to startups and only to “significant platforms”. He was reacting to an Economic Times story published Monday on how start-ups are concerned about the government’s missive to get permission before launching new AI models. Chandrashekhar on microblogging site X said, “Recent advisory of @GoI_MeitY needs to be understood. Advisory is aimed at the Significant platforms and permission seeking from Meity is only for large platforms and will not apply to startups.”Elevate Your Tech Prowess with High-Value Skill CoursesOffering CollegeCourseWebsiteIIM LucknowIIML Executive Programme in FinTech, Banking & Applied Risk ManagementVisitIndian School of BusinessISB Product ManagementVisitIIT DelhiIITD Certificate Programme in Data Science & Machine LearningVisitSeveral AI startups had raised concerns over the government advisory arguing that it is anti-innovation and not forward looking. The minister said that the advisory is aimed at untested AI platforms from deploying on the Indian Internet. The process of seeking permission, labelling and consent-based disclosure to users about untested platforms is “an insurance policy” to platforms who can otherwise be sued by consumers, he said. The government’s advisory though was slammed internationally by global AI experts. Bindu Reddy, chief executive of Abacus.ai, which uses Gen AI to build Applied AI and LLM agents and systems at scale, remarked on X, “India just kissed its future goodbye!”Discover the stories of your interestBlockchain5 StoriesCyber-safety7 StoriesFintech9 StoriesE-comm9 StoriesML8 StoriesEdtech6 Stories”Every company deploying a GenAI model now requires approval from the Indian government! That is, you now need approval for merely deploying a 7b open source model,” she said. “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!” she exclaimed. “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,” she rued. Petro Domingos, professor of computer science at the University of Washington said on X, “The Indian government takes the lead in the race to the stupidity singularity.”Aravind Srinivas, chief executive of Perplexity, which has build an answer engine powered by AI, said on X, “Bad move by India”.Each generative AI and Large Language Model industry is in a hyper-growth phase right now, Shorthills AI co-founder Paramdeep Singh told ET. Large organisations and open-source communities are releasing models and products at a break-neck speed. While organisations struggle to get to market as fast as possible to gain the first mover advantage, they are bound to release products that could have some glitches, he said. Apart from this, it is a fact that AI models are known to hallucinate and give out incorrect information when asked about facts, he explained. The way these models are trained, they are more of predictors of next characters and words rather than true search engines for information, he said. It is going to be extremely hard to test these models for all possible questions and information searches, he argued. Organizations are putting efforts in the right direction to improve the results provided by AI models, he said. Some of these techniques like Read and Retrieve (R&R), Retrieval Augmented Generation (RAG), and Reinforcement Learning from Human Feedback (RLHF) put guardrails around what the model returns. But this is an industry in the nascent stage and it would take time to get to full maturity, he argued.It is the government’s job to make sure that there is no mass spread of misinformation, Singh said. The government’s advisory on March 1 said, “All intermediaries or platforms to ensure that use of Artificial Intelligence model(s) /LLM/Generative AI, software(s) or algorithm(s) on or through its computer resource does not permit its users to host, display, upload, modify, publish, transmit, store, update or share any unlawful content as outlined in the Rule 3(1)(b) of the IT Rules or violate any other provision of the IT Act.”
