Despite AI’s potential to address societal challenges, it also presents risks like deepfakes and misinformation
The Indian government is contemplating the creation of a national Artificial Intelligence Safety Institute (AISI), reflecting its growing focus on AI governance and safety. At the *Building AI Companions for India* event hosted by Microsoft in Bengaluru, S Krishnan, Secretary at the Ministry of Electronics and Information Technology (MeitY), discussed this initiative alongside Mustafa Suleyman, CEO of Microsoft AI.
“I think AI Safety Institute (AISI) seems to be the flavour of the day across the world, and we are in the process of trying to establish one ourselves in order to understand this better,” Krishnan remarked, highlighting the government’s intent to join the global movement for ensuring safe and ethical AI development.
The proposal for an AISI has been under discussion for several months. Last month, MeitY convened a meeting to outline the institute’s objectives, budget, and operational framework. Krishnan underscored the importance of a proactive regulatory approach, stating, “I think, you know, we sort of waited for things to go wrong this time around. I think it’s time to be really thoughtful and deliberate and not treat that as such a taboo.”
Despite AI’s potential to address societal challenges, it also presents risks like deepfakes and misinformation. Krishnan acknowledged that India’s existing laws have been moderately effective in countering these threats, particularly during the 2024 elections. “With things like misrepresentation and deepfakes, we feared so much both in the Indian election and in the other elections held throughout the world in 2024, but I think existing legislation has proved reasonably effective in addressing those issues,” he said.
Mustafa Suleyman highlighted the uncertainties surrounding AI’s evolution, particularly the risks posed by advanced models capable of self-improvement. “The tricky thing for us to figure out in the next five years is when a model starts to have the ability to improve itself independently, those kinds of recursive self-improvement mechanics. You know, we sort of don’t really know exactly how they’re going to turn out,” he said, stressing the need for careful policy interventions as AI capabilities expand.
Globally, the trend of establishing AI safety institutes has gained momentum. The United Kingdom pioneered this effort in 2023 by launching its AI Safety Institute, which focuses on testing and mitigating risks associated with advanced AI systems. Major tech companies, including OpenAI, Meta, Google DeepMind, and Microsoft, signed voluntary agreements to provide early access to their models for safety testing. Ian Hogarth, chair of the UK Government’s AI Foundation Model Taskforce, noted, “We hosted the Bletchley AI safety summit… and we got an agreement between China, the US, and Europe across these risks.”
Following the UK’s lead, the United States established its AI Safety Institute, securing agreements with firms like OpenAI and Anthropic. Elizabeth Kelly, director of the US institute, emphasised its role in fostering innovation through trust and safety. “Safety promotes trust, which promotes adoption, which drives innovation, and that’s what we are trying to promote at the US AI Safety Institute,” she said.
As India explores this initiative, the establishment of an AISI could place the country at the forefront of AI safety and governance. By fostering collaboration, advancing safety research, and aligning with global standards, the institute has the potential to ensure that AI technologies develop in ways that benefit society while mitigating risks.
