AI accelerating science, augmenting human capability: Google DeepMind’s Manish Gupta

Artificial Intelligence has emerged as humanity's most powerful force for progress, Manish Gupta, Senior Director of Google DeepMind said on Tuesday, adding that AI's evolution will drive significant shifts including accelerating scientific discovery and augmenting human capabilities.

The senior executive at Google DeepMind, the tech giant's premier AI research lab, emphasised that India is at the forefront of utilising Artificial Intelligence to solve real-world, tangible problems.

Google, the architect of AI models, including Gemini and Gemma that are integrated for search enhancements, cloud services, and enterprise tools, has recently launched Gemini 2.5 Pro Preview, which excels in coding and multimodal tasks.

At DevSparks 2025 in May, Gupta highlighted that DeepMind’s AlphaFold has predicted the structures of all 200 million known proteins and is now widely used for drug discovery and designing enzymes for biodegradable plastics, compressing what once took years of PhD research into seconds.

On Tuesday, the company announced funding support of $8 million to India's AI Centers of Excellence for health, agriculture, education, and sustainable cities, and committed $400,000 to support development of India's health foundation model.

The tech giant is also supporting Gnani.AI, CoRover.AI, and BharatGen with $50,000 grants each for building models serving Indic language solutions.

As such, Google's strategic focus on India was solidified in October with its announcement of record $15 billion investment that would go into building an AI infrastructure hub in Andhra Pradesh, which would include a gigawatt-scale data centre in partnership with Adani Group.

"We see AI as a very powerful force for humanity, and we are already seeing it dramatically power and accelerate scientific discovery. We are also seeing it become a very powerful augmentation for human capability and we are seeing India at the forefront of using AI to solve real world problems," Gupta told PTI.

The true impact of AI, he explained, lies in tackling `root node problems', grand challenges whose solutions can unlock massive, cascading benefits for humanity, he noted.

The impact is not limited to life sciences, alone, but covers anything from material science to clean energy, education to agriculture. Beyond scientific discovery, AI is shifting from predictive and conversational capabilities to multi-modal reasoning and agentic capabilities.

AI models can now interact like humans, through text, voice, cameras, and video, allowing them to not just converse, but to initiate action on the user’s behalf.

"So conventionally we've thought about predictive capabilities and kind of conversational capabilities. But now we are moving beyond that to a multi-modal reasoning. These AI models are able to interact with you the way we humans interact through not just text, but voice, cameras, through videos," he said.

The powerful reasoning capabilities and agentic capabilities means AI model today "can go beyond just a conversation, and can actually take action on the user's behalf".

Google on Tuesday marked its commitment to bolster India's AI ecosystem, with a slew of new collaborations and funding announcements. It is providing Wadhwani AI with $4.5 million in funding to support multilingual AI-powered applications for health and agriculture; and partnering with ReNew Energy to support new 150 MW solar project in Rajasthan and reduce emissions across value chain.

Gupta also highlighted Google's work on inclusive AI in India, including Project Vaani in partnership with IISc Bangalore, which collects speech data from 773 districts to improve AI understanding of Indian languages. The first phase has released data covering 59 Indian languages, including 15 previously zero-corpus languages.

Google announced funding of $400,000 to support new collaborations that will leverage 'MedGemma' to build India's Health Foundation Models, according to the company release.

Such models aim to raise the efficiency of healthcare providers and improve patient outcomes across India.

Google said Ajna Lens will work with experts from the All India Institute of Medical Sciences (AIIMS) to build models that will support India-specific use cases in Dermatology and OPD Triaging.

The resulting models, it said, will contribute to India's Digital Public Infrastructure and their outcomes will be made accessible to the ecosystem. Additionally, researchers, AI experts, and clinicians from IISc will explore using AI models for broader clinical applications.

To push its inclusive AI agenda, Google announced a $2 million founding contribution to establish the new Indic Language Technologies Research Hub at IIT Bombay. The initiative aims to ensure global advancements serve India's linguistic diversity, Google said.

Earlier this year, DeepMind unveiled Genie 3 (August 2025), a general-purpose AI model for generating interactive 3D environments in real time, highlighting the lab’s focus on multi-modal reasoning and simulation capabilities for research and training applications.

The company is also focusing on building energy-efficient models using the Matryoshka architecture, exemplified by Gemini Nano and Gemini Flash, which adjust size dynamically and deliver higher computational efficiency per watt than comparable models.

(With additional inputs from PTI)