Google announced its Safety Charter for India, emphasizing its efforts to leverage artificial intelligence (AI) technology to combat cybercrime across various platforms. As India’s digital economy continues to expand, the necessity for trust-based systems has become increasingly critical. The tech firm is utilizing AI in its products, nationwide initiatives, and efforts to identify and mitigate vulnerabilities in enterprise software, while also advocating for the responsible development of AI technologies.
Significant Achievements Under Google’s Safety Charter for India
In a recent blog entry, Google outlined its progress in effectively identifying and curtailing online fraud and scams within its consumer products and enterprise software. The company referenced a report indicating that fraud linked to UPI (Unified Payments Interface) resulted in losses exceeding Rs. 1,087 crore in 2024, while overall financial damages from unregulated cybercrimes reached an alarming Rs. 20,000 crore in 2025.
Additionally, Google noted that malicious entities are increasingly employing AI to refine their cybercriminal tactics. Techniques such as AI-generated content, deepfakes, and voice cloning have emerged as tools used for convincing fraudulent schemes.
To enhance cybersecurity measures in India, Google is integrating its policies and security technologies with the DigiKavach initiative. Furthermore, the company has formed a partnership with the Indian Cyber Crime Coordination Centre (14C) to bolster user education on cybercrimes in a phased strategy over the coming months.
In terms of its accomplishments in this domain, Google reported the removal of 247 million advertisements and the suspension of 2.9 million fraudulent accounts that violated its guidelines, adhering to local and national regulations.
When it comes to Google Search, the company claims that it is utilizing AI models to identify 20 times more fraudulent web pages before they can appear in search results. Additionally, it has significantly decreased the instances of scam websites impersonating customer service and government entities by more than 80 percent and 70 percent, respectively.
Google Message has recently integrated a new AI-driven Scam Detection feature, with the company asserting that this security tool flags over 500 million suspicious messages each month. Users receive warnings when clicking on URLs shared by unknown contacts, with the alert displayed more than 2.5 billion times to date.
The Google Play Store has reportedly blocked almost 60 million attempts to install high-risk applications, which included over 220,000 unique apps targeting more than 13 million devices. The Google Pay app has also issued alerts regarding potential scams on 41 million transactions identified as suspicious.
Broadening its focus, Google is also prioritizing the protection of its enterprise products against potential cybersecurity threats. It launched Project Zero in collaboration with DeepMind to uncover previously unidentified vulnerabilities in popular enterprise software, such as SQLite. In the case of the SQLite flaw, an AI agent was employed to detect the issue.
In addition, Google is partnering with IIT Madras to explore Post-Quantum Cryptography (PQC), which consists of cryptographic algorithms designed to safeguard systems against the forthcoming threats posed by quantum computing. These algorithms are vital for encryption, digital signatures, and secure key exchanges.
On the topic of responsible AI usage, Google asserted that its models and infrastructure undergo comprehensive testing to guard against adversarial attacks using both internal assessments and AI-assisted red teaming methods.
For precise labeling and accuracy of AI-generated content, Google is implementing SynthID to embed an unobtrusive watermark in text, audio, video, and images produced by its models. Additionally, the company requires YouTube creators to clearly disclose AI-generated content. The Gemini chatbot also features a double-check capability, enabling users to verify inaccuracies via a Google Search.
