Google launches safety charter to secure India’s AI-driven digital transformation

Google has released its Safety Charter for India’s AI-led Transformation, outlining a comprehensive framework aimed at enhancing security and trust within India’s rapidly evolving digital landscape. As artificial intelligence (AI) becomes increasingly embedded in services and infrastructure, the charter highlights Google's commitment to safeguarding users, businesses, and government institutions.

The company noted that India’s digital economy has thrived due to improved accessibility, affordable devices, and widespread adoption of digital services. However, it stressed that user trust, the bedrock of this growth, must be carefully protected. Google emphasised the importance of responsible technology stewardship, especially as AI continues to drive innovation while also introducing more complex cybersecurity threats.

“The internet can serve as a powerful growth engine, so long as citizens feel safe engaging with it,” wrote Heather Adkins, VP of Security Engineering, and Preeti Lobana, VP and Country Manager for India, in a Google blog post. “As responsible stewards of India’s tech ecosystem, it is vital that we preserve user trust as AI becomes more deeply integrated into our digital lives.”

The Safety Charter is built on three core pillars: responsible AI, cybersecurity, and user safety. Under these themes, the blog detailed how AI is being used to strengthen software security, how Google’s investment in products and programmes is protecting consumers, and how AI is helping close the gap between cyber attackers and defenders. “Consider it our roadmap for addressing the emerging challenges of the internet age, in collaboration with the broader ecosystem,” the post explained.

Tackling online threats and financial fraud

A key focus of the charter is protecting users from increasingly sophisticated online scams, particularly those driven by deepfakes, voice cloning, and AI-generated content. Industry estimates suggest that financial fraud, especially involving the Unified Payments Interface (UPI), cost Indians over Rs 1,087 crore in 2024, with potential losses projected to reach Rs 20,000 crore by 2025 if left unaddressed.

As part of its response, Google has expanded its DigiKavach initiative, which has reached over 177 million people through a combination of user awareness campaigns and in-product safety features. The company also announced a partnership with the Ministry of Home Affairs’ Indian Cyber Crime Coordination Centre (I4C) to scale up national awareness efforts on cybercrime in the coming months.

In 2024 alone, Google suspended 2.9 million accounts for policy violations, including those linked to financial fraud, and removed over 247 million misleading ads. Globally, AI improvements have enabled Google Search to detect 20 times more scam-related webpages, including those generated by large language models. Targeted protections have also led to a 70% reduction in impersonation attacks on government portals and an 80% drop on customer service platforms.

Mobile and platform-level safety enhancements

On Android, Google delivers more than 2.5 billion monthly warnings about suspicious links from unknown senders, while its AI-powered Scam Detection blocks over 500 million dubious messages in Google Messages, all processed on-device to maintain user privacy.

Meanwhile, Google Play Protect scans over 100 billion installed apps daily across billions of devices. Since launching a pilot programme in India in late 2024, more than 220,000 harmful apps have been blocked on over 13 million devices, preventing around 60 million attempts to install high-risk applications. In the same year, Google Pay issued over 41 million scam-related transaction warnings to Indian users.

Strengthening public infrastructure and cyber education

Google reiterated its "secure-by-design" and "secure-by-default" approach while enhancing cybersecurity support for public and enterprise infrastructure. Through Mandiant-led threat intelligence and Google Cloud’s M-Trends report, organisations are being kept informed about evolving threats such as ransomware and cloud breaches. Notably, a collaboration with DeepMind enabled the first AI system to identify a real-world memory safety flaw in the SQLite database.

To further support cybersecurity education and capacity building, Google.org announced an additional USD 5 million commitment to The Asia Foundation, on top of a previous USD 15 million. This funding will help expand the APAC Cybersecurity Fund, supporting students and micro, small, and medium enterprises (MSMEs) across India. It will also fund new cybersecurity clinics in partnership with Indian universities. Additionally, Google is working with IIT Madras on research into Post-Quantum Cryptography (PQC).

Google also reaffirmed its efforts to enhance transparency around AI-generated content. The company’s SynthID technology, which embeds invisible digital watermarks into AI-generated material, has now been used to mark over 10 billion pieces of content. The company requires creators to label synthetic images in Google Search and to disclose AI-generated content on YouTube. “We are investing in tools to detect AI-generated content,” the June 17, 2025 blog post stated. “These safeguards are essential to maintaining user trust and ensuring the integrity of the information ecosystem."