AI browsers could leave users penniless: A prompt injection warning

The rise of AI-powered and agentic browsers introduces a new class of cybersecurity threats—prompt injection attacks. These attacks exploit the language-processing capabilities of LLMs embedded in browsers.