Google has issued a warning after detecting a new form of cyber attack that specifically targets Gmail users by exploiting artificial intelligence systems. The attack, known as indirect prompt injection, tricks AI-powered features into revealing sensitive information or performing unintended actions.

According to Google’s cybersecurity team, the threat works by inserting malicious instructions into seemingly harmless content, such as emails or links, which then manipulate AI tools to compromise user data. Experts have cautioned that this method is harder to detect than traditional phishing or malware, making it a significant risk.

The company has advised Gmail users to remain cautious when interacting with unfamiliar emails or attachments, even if they appear genuine. Google has also strengthened its security measures to detect and block such attacks before they reach users.

Cybersecurity specialists note that the rise of AI-driven threats highlights the urgent need for advanced safeguards and awareness among internet users.