Reportedly, a researcher recently discovered a security flaw in Gmail's AI-generated summaries that could allow threat actors ...

The exploit, known as a prompt injection attack, evades detection by reducing the prompt font size and changing it to white to blend in.