Editor’s note: This article is based on insights from a podcast series. The views expressed in the podcast reflect the speakers’ perspectives and do not necessarily represent those of this publication. Readers are encouraged to explore the full podcast for additional context.
In an episode of the “CAIO Connect” podcast, host Sanjay Puri interviewed Owen Larter, who is in charge of global policy at Google DeepMind, at the AI Innovation Impact Summit. The interview revolved around the defining question of our era: How can we scale AI responsibly while making sure it is trusted, inclusive, and safe?
Right from the beginning, Larter stressed that the future of AI not only relies on technological innovation but also on the creation of a robust trust framework. At Google DeepMind, he said, the development of more sophisticated AI systems is inextricably linked with collaboration with governments to help them understand the technology and manage it effectively. AI systems cannot be “black boxes” that are deployed at a large scale; they have to be tested and improved.
India, he said, is also positioning itself as a significant player in AI. The country’s investments in digital public infrastructure and connectivity provide a strong foundation for the adoption of AI. He shared, “India is one of our most important markets, so it’s great to see people here using Gemini. It’s great to see people here using important systems like CoScientist, AlphaFold.”
Projects that apply AI to agriculture, education, and scientific research show that this is not just a Silicon Valley phenomenon. AI is increasingly being used as a tool for farmers to track crop health, students to improve learning outcomes, and researchers to speed up medical discoveries.
READ: Prith Banerjee on making ‘physical AI’ real for global South (May 15, 2026)
A case in point is the increasing adoption of sophisticated AI technology by Indian researchers. Thousands of researchers are utilizing protein prediction software to gain a better understanding of diseases and work on novel treatments. This is a further example of a point made: the transformative power of AI will be realized in all industries, not only in tech centers.
However, scale by itself is not sufficient. Accessibility and inclusion are still core issues. The linguistic diversity of India is both an opportunity and a challenge. Larter pointed out the efforts being made to enhance the performance of AI in the Indic languages so that the AI is useful regardless of whether the person speaks Tamil, Gujarati, Hindi, or Telugu. Multilinguality is not just a functionality but the basis for accessibility of AI.
“We really want this technology to be used right across society for all kinds of reasons. It’s the right thing to do. It also makes sense from a business perspective. We want these to be good products that are useful to people in whatever language they’re speaking…. We’ve also made additional investments recently, working with the IIT Bombay to study Indic languages and how we can improve the performance of AI systems across languages in India and more broadly,” Larter shared.
Another important theme was the emergence of agentic AI, which has the ability to function more independently and perform tasks on behalf of users. Although such abilities offer unprecedented opportunities for productivity, they also raise new issues of governance. How can we test agentic systems? How can we make them secure? How can we develop common standards to give firms and citizens assurance?
READ: Anu Bradford on why India should build a hybrid AI governance model (May 11, 2026)
Larter made a very pertinent comparison to the early days of the internet, where open standards had facilitated its growth and trust. By the same token, the “agentic economy” will also need strong safety nets and global collaboration.
“This is not a transition that any one group is going to be able to manage by themselves. Industry, government, civil society, academia, we need to share expertise and perspectives across these groups so that we’re building technology that is useful and also safe and secure and trustworthy,” he said.
One thing that was clear throughout the discussion on the CAIO Connect Podcast is that governments need to take the lead. They cannot just sit back and watch as AI develops.
Larter explained, “I do think that governments are going to play a really critical role here as part of this transition. I think they will need to continue to lead and use the technology themselves. I think we need to make sure that AI is not just something that is being used outside of government, but that governments need to be right at the forefront of how they use this technology.”
As Larter concluded, the process of transitioning into an AI-enabled future is a collective process. Innovation in the industry needs to be complemented by leadership in the public domain. Progress needs to be accompanied by effective governance. And access needs to extend beyond experts and elites.
The age of AI has arrived. The question is whether we will build it on the foundation of trust.