India well-positioned to help shape future of AI: Google DeepMind's Jeff Dean

(R-L) Jeff Dean, Grace Chung and Manish Gupta at Research@ event

India’s large pool of technical talent positions the country well to help shape the future of artificial intelligence, Google's AI head Jeff Dean has told Moneycontrol.

"India has tremendous technical talent. There is a long tradition of computer science students in India doing amazing things" Dean said at the company's first Research@ event in India held in Bengaluru on February 1.

Dean, who serves as the chief scientist at Google DeepMind and Google Research, said anyone with an interest in computer science is turning their attention to machine learning and what it can do.

India is well represented in this area since the country has many students and industry executives already in the field of machine learning and artificial intelligence, he said. "You see this in contributions to top conferences. There's many people with Indian backgrounds there," he said.

Dean was the co-lead of Alphabet's Gemini project along with Oriol Vinyals, vice president of research and deep learning lead at Google DeepMind.

Gemini is the first AI model from Alphabet after the merger of its AI research units, DeepMind and Google Brain, into Google DeepMind, led by DeepMind CEO Demis Hassabis.

Also read: Google unveils Gemini, its largest AI model, to take on OpenAI

Gemini has been built from the ground up and is "multimodal" in nature, meaning it can understand and work with different types of information, including text, code, audio, image and video, at the same time. It can also run on everything ranging from mobile devices to data centres.

Dean said the industry needs to move to a more incremental learning system where everyone is continuously working to improve a single AI model

"If we could have a system where everyone was collectively improving the same model, you could incrementally extend the capabilities of the model," he said.

Dean also mentioned that he cares "very deeply" about the ability of AI models to work across many languages.

"One of the efforts I helped encourage was our 1,000 language effort, where we want to make the models that we are training work across the 1,000 most commonly spoken languages in the world," he said.

The initiative was first announced by Google in November 2022.

Also read: Google parent to make 'meaningful' investments to double down on its AI efforts, says CEO Sundar Pichai

Dean said it's crucial to make these developing technologies, sometimes launched first in English or a few languages, accessible to speakers of various languages across different communities.

"The more accessible we make these technologies to people, the better off we will be...we will provide access to information in ways that previously have been underserved. That's going to be really important" he said.

Dean noted that many languages don't have much written material but are still heavily spoken by tens of millions of people. "It's a very important problem in India, because there's so many languages spoken in this country," he said.