AI Safety: Google’s Commitment to Safe and Trusted Artificial Intelligence

AI safety is being prioritised as essential infrastructure for Artificial Intelligence to realise its significant transformative potential. This was the key theme of the ‘Safe and Trusted AI’ event organised by Google in the capital, leading up to the AI Impact Summit 2026. The company is particularly focused on safeguarding children, teenagers, and the elderly in its AI safety initiatives.

Insights from Google’s Experts

Evan Kotsovinos, Vice President for Privacy, Safety, and Security at Google, remarked that safety serves as the crucial infrastructure that enables AI to confidently enter sectors such as finance, healthcare, and education. This foundation is vital for stimulating economic growth. He underscored that in India, where a thriving digital economy is evident, Google is developing AI systems aimed at sustaining user trust during the country’s advancement towards AI integration. Kotsovinos noted that AI-driven defences operating on devices provide substantial protection against malicious entities targeting individuals and critical public services. He indicated that India’s vast and diverse market positions it as a global testing ground for reliable AI safety, asserting that India will lead the way to equitable AI solutions for the entire Global South.

Comprehensive Safety Strategy

Preeti Lobana, Country Manager for Google India, elaborated on the company’s “360-degree” approach to safety. This strategy combines on-product and cloud-based protections, fosters digital literacy, continually invests in safety research, and collaborates with the broader ecosystem. She encapsulated the core objective as fostering an environment where safety is inherent to every aspect of technology.

Prioritising Protection for Vulnerable Groups

Google places significant importance on protecting children, teens, and the elderly across all its platforms. The company is implementing safeguards and features that function directly on devices, making them quicker, more private, and perpetually active, while also investing in digital literacy initiatives to close existing knowledge gaps.

Innovative Features for User Safety

This initiative is powered by Gemini Nano and is being introduced on Pixel phones. It provides real-time analysis of calls, identifying potential scams entirely on the device without recording audio, transcripts, or sharing information with Google. The feature remains off by default, applies solely to calls from unknown numbers, and users can deactivate it at any time.

This new pilot feature will issue critical alerts when users are deemed to be at elevated risk. If a user shares their screen while talking to an unknown caller and accesses financial applications such as Google Pay, Navi, or Paytm, an immediate warning will appear on the device. The user can then terminate the screen shared session and the call with a single tap. This feature is specifically crafted to thwart scams that target UPI and banking applications during live conversations.

Advancing AI Safety Research and Education

Google is expanding the reach of its SynthID Detector and is launching the SynthID text watermarking tool as an open-source component of its Responsible GenAI Toolkit. This watermarking capability aids academic institutions, researchers, and publishers in identifying AI-generated images and audio. Early access to SynthID Detector and API is being granted to strategic partners, including educational institutions, researchers, and publishers such as Jagran and PTI.

Investment in Digital Literacy Initiatives

In addition to technical advancements, Google is heavily investing in promoting digital literacy and enhancing the wider technology ecosystem. The flagship Learn and Explore Online (LEO) programme will be introduced in India in December 2025, aiming to educate teachers, practitioners, and parents on utilising parental tools and creating age-appropriate online experiences. Google is contributing a grant of $200,000 USD from Google.Org to the CyberPeace Foundation in India, supporting capabilities development, delivering AI-driven cyber-defence tools, and cultivating safer digital learning environments for children and teenagers. The “Sach Ke Sathi, DigiKavach for Seniors” programme aims to provide in-person training to over 5,000 seniors across approximately 30 cities.

Collaborations to Enhance Safety

In collaboration with the Reserve Bank of India (RBI), Google has facilitated the publication of an authenticated list of approved digital lending applications and their associated NBFCs, assisting users in recognising trustworthy apps. The company is also working with IIT Madras and CeRAI (Centre for Responsible AI) to propel AI safety research, reflecting the diverse language and device landscape of India.

Innovations in Security Technology

Google has introduced CodeMender, a self-defending security system that autonomously detects and fixes zero-day vulnerabilities in open-source software. Recent innovations like Private AI Compute, Parfait, and VaultGemma align with Privacy Enhancing Technologies (PETs), designed to assist the broader ecosystem in creating privacy-preserving AI systems. Furthermore, the company is developing Enhanced Phone Number Verification (ePNV), a new technology intended to replace less secure SMS-based one-time passwords with a SIM-based verification method to bolster sign-in security.