Hidden Prompt Injection Vulnerability Remains in Gmail’s AI Summaries After Google Mitigations

Security experts warn that Google’s filters have not fully stopped hidden instructions from enabling attackers to trigger fake phishing alerts through AI summaries.