Google's Gemini Live takes center stage at Google for India 2024

Today, at the 10th edition of Google for India 2024, Gemini Live has taken centre stage at the event. Hema Budaraju, director of product management at Google shared that India is leading the way for many of Google's innovations.

She added that starting today (October 3, 2024), Gemini is available in Hindi and eight more Indian languages. “We believe this can be a game changer for teachers, entrepreneurs, artists, everyone, and we are making it even more accessible with the power of your voice in India, we know that it's one of the most popular ways that people use technology. Over 40% of Gemini Indian language users here are already using their voice to get quick inspiration," said Budaraju.

Moving further in her session, she also announced that the company has also introduced updates in Google Maps. If someone wants to look for restaurants in their city, Gemini will summarise the location and reviews on Google Maps in seconds. It will also help in making users' journeys safe by introducing two new features of real-time weather alerts. If it is a foggy morning or a continuous rain, users will now be able to report the same and be alerted about these disruptions.

Moreover, the tech giant is also adding to its AI overhaul of search. Google is now introducing video search, where one can film a problem they want to resolve (like seeing their TV screen flicker unusually) and the Google search's AI Overview will take a look.

"We know it's sometimes hard to explain what you need with just words. Soon you'll be able to ask questions with a video. Let's look at one problem we commonly face. Now you can take a quick video and ask out loud, how do I fix this? Search will make sense of the video and the question to give you a helpful AI overview,” added Budaraju.

Building upon these announcements, other key speakers at the event highlighted Google's ambitious plans for AI in India across various sectors.

Bikram Bedi, managing director at Google Cloud India, unveiled an open-source AI agent framework for backend-enabled open networks, powered by Gemini models. This framework aims to streamline transactions and discovery across multiple sectors, including agriculture, education, and job searching.

He emphasised, "We are bringing the power of backend enabled open networks and our Gemini, an open source AI agent framework for all backend enabled open networks engineered using Gemini models will now be available to every Indian business and every user."

Furthermore, Bedi announced plans to launch data storage and machine learning processing capabilities for Gemini 1.5 entirely within India later this year, addressing data sovereignty concerns for sensitive sectors.

Interestingly, Google is also partnering with the D8 step Foundation to launch 'DPI in a box', a plug-and-play model based on open networks that can help nations build their own digital infrastructure.

Shekhar Khosla, VP of marketing at Google India, focussed on the critical aspect of AI skilling. He unveiled the 'AI Skills House' programme, designed to empower 10 million Indians with AI literacy.

"This initiative is designed to tap into India's talent pool and enable AI literacy for students, job seekers, educators, developers and civil officials", he stated. The programme will offer free online courses, initially in English, with plans to expand to seven Indian languages.

Google.org, the company's philanthropic arm, is supporting this initiative with a new $4 million grant to the Central Square Foundation for a nationwide AI literacy and awareness initiative. Additionally, Google.org is launching a fellowship to support Rocket Learning in developing a Gen AI-powered tutor for young children across India.

Manish Gupta, director of Google Research India, highlighted the company's progress in AI research and development. He presented the Indic GenAI benchmark, covering 29 Indian languages, to address the gap between AI capabilities in English versus Indic languages. Gupta also announced advancements in multilingual AI, stating, "We're also working towards making our large language models understand hundreds of languages using an enhanced version of PaLM 2 model".

Gupta further revealed that Google's agricultural landscape understanding model, which utilises satellite imagery for farm-level insights, is now ready for developers to use. "We hope this technology will go on to benefit millions of farmers across the country", he added.