\n\tThe approach marks a shift in how Google DeepMind\u2019s India teams work, using AI not only to build software but to help chart a path toward artificial general intelligence (AGI), which can perform human-like tasks across domains. Demis Hassabis, Google DeepMind\u2019s chief executive officer who shared the 2024 Nobel Prize in Chemistry, has said that AGI could arrive as early as 2029.<\/p>\n
\u00a0<\/p>\n
\n\tBy assigning routine testing to AI agents, DeepMind hopes to speed up specialised applications in drug discovery and materials science, while strengthening its foothold against OpenAI and Anthropic in India\u2019s developer market.<\/p>\n
\u00a0<\/p>\n
\n\tSpeaking at the India AI Impact Summit in February, Hassabis proposed an \u201cEinstein Test\u201d for AGI: Training a model exclusively on pre-1911 scientific data to see if it can independently derive general relativity.<\/p>\n
\u00a0<\/p>\n
\n\tAI is a powerful research tool, Manish Gupta, senior director for India and APAC at Google DeepMind, and Seshu Ajjarapu, the company\u2019s senior director of applied AI, told Business Standard in an interview in Bengaluru.<\/p>\n
\u00a0<\/p>\n
\n\tGoogle DeepMind teams, including those in India, are using AI agents to improve Gemini by identifying capability gaps, curating data and verifying improvements, said Gupta. Teams in Bengaluru are improving Gemini\u2019s reasoning and multilingual performance. Automating parts of AI research gives scientists powerful tools for complex challenges such as disease discovery and new materials.<\/p>\n
\u00a0<\/p>\n
\n\t\u201cAs we move towards more powerful AGI capabilities, we have the opportunity to tackle virtually every disease, develop cures, address renewable energy and solve many of the toughest problems that remain unsolved,\u201d Gupta said. \u201cWe have that opportunity to make much faster progress on many of those very tough problems. That work is being done globally, including in India.\u201d<\/p>\n
\u00a0<\/p>\n
\n\tAjjarapu said AGI\u2019s best value will lie in its ability to benefit humanity by accelerating scientific discovery, including advances in medicine, biology, climate and energy.<\/p>\n
\u00a0<\/p>\n
\n\tResearchers in India have access to DeepMind AI tools such as AlphaGenome, which analyses the impact of genetic mutations, and AI Co-Scientist, a virtual scientific collaborator.<\/p>\n
\u00a0<\/p>\n
\n\tAcross India, new and expectant mothers are adopting life-saving health practices through programmes run by the nonprofit Armman. The initiative powers its outreach using machine learning models developed with pro bono support from Google DeepMind researchers.<\/p>\n
\u00a0<\/p>\n
\n\tMore than 150,000 Indian researchers are leveraging the AlphaFold Protein Structure Database for open-access scientific discovery. They are deploying a DeepMind tool called AlphaFold to address complex health crises, from autoimmune diseases to cancer.<\/p>\n
\u00a0<\/p>\n
\n\tResearchers at All India Institute of Medical Sciences (AIIMS) in New Delhi are using Google’s open-source MedGemma models to build localised tools for leprosy and reproductive health. Google plans to release these models to India’s broader developer ecosystem.\u00a0<\/p>\n
\n\tGupta said that Indian researchers are fine-tuning these models with local datasets \u2014 noting that pairing global pre-trained models with targeted Indian data yields far more accurate systems than building them from scratch locally.<\/p>\n
\u00a0<\/p>\n
\n\tGemini Live, a conversational voice feature for Google\u2019s AI, now supports 25 Indian languages and dialects, including Sanskrit, Bhojpuri and Maithili, letting users interact with the assistant in their native language. Gupta said such models have significantly narrowed the performance gap between English and Indian languages, although improvements are needed.<\/p>\n
\u00a0<\/p>\n
\n\tGoogle worked with the Indian Institute of Science, Bengaluru, to create Project Vaani, one of the largest datasets of Indian dialects.\u00a0<\/p>\n
\n\tGoogle DeepMind and the Partnerships Innovation team have launched the Agricultural Monitoring and Event Detection API to enhance farm oversight across India. Powered by machine learning, satellite imagery and three years of historical data, the tool identifies crop types, field sizes and planting dates.\u00a0<\/p>\n
\n\tComplementing this is Earth AI \u2014 a suite of Gemini-powered models designed for environmental tracking and disaster response. Together, these tools deliver data-driven insights to refine farm management, address soil and climate needs, and forecast harvest yields.<\/p>\n
\u00a0<\/p>\n
\n\tGupta said he views AI less as a geopolitical race and more as a way to address India’s challenges in health care, education and agriculture. He noted that India\u2019s growing developer ecosystem and startups such as Sarvam are building foundational AI models. He said India should now apply AI to fundamental scientific challenges, such as drug discovery and sustainable energy.<\/p>\n
\u00a0<\/p>\n
\n\tAsked how Google DeepMind differentiates itself from rivals such as OpenAI, Anthropic, Meta and Microsoft in winning Indian developers and enterprises, Ajjarapu said the company’s advantage lies in its end-to-end AI stack.<\/p>\n
\u00a0<\/p>\n
\n\tAjjarapu noted that Google builds across the entire AI stack \u2014 from custom infrastructure like Tensor Processing Units to foundational models, open systems like Gemma, specialised tools such as MedGemma, and enterprise software.<\/p>\n
\u00a0<\/p>\n
\n\tRather than chasing labels like AGI or artificial superintelligence, Ajjarapu evaluates progress through practical benchmarks \u2014 specifically the real-world problems solved and how AI is delivered. While systems have surpassed human capability in specialised domains, he noted that true general intelligence will take longer and requires strict oversight. \u201cWe need to shepherd this technology carefully with a lot of responsibility and clear markers,\u201d he said.<\/p>\n
\u00a0<\/p>\n","protected":false},"excerpt":{"rendered":" The approach marks a shift in how Google DeepMind\u2019s India teams work, using AI not only to…\n","protected":false},"author":2,"featured_media":127313,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[4],"tags":[6744,405,41038,11553,63747,3013,7718,63750,3234,7543,63751,55241,63748,63749,63746],"class_list":["post-127312","post","type-post","status-publish","format-standard","has-post-thumbnail","category-agi","tag-agi","tag-ai-agents","tag-ai-co-scientist","tag-alphafold","tag-alphagenome","tag-artificial-general-intelligence","tag-demis-hassabis","tag-earth-ai","tag-gemini-ai","tag-google-deepmind","tag-indian-ai-developers","tag-manish-gupta","tag-medgemma","tag-project-vaani","tag-seshu-ajjarapu"],"_links":{"self":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts\/127312","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/comments?post=127312"}],"version-history":[{"count":0,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts\/127312\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/media\/127313"}],"wp:attachment":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/media?parent=127312"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/categories?post=127312"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/tags?post=127312"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}