Insights on DLP & Security

White text on white background, instructions buried in file metadata, text set at 1-point font — these are the tools job applicants are using to manipulate AI in corporate HR systems. This is not fiction. It is a documented attack affecting companies worldwide.

Every day, millions of employees copy sensitive data into ChatGPT, Clipchamp, Gemini and other AI tools. They are not doing it maliciously — they simply do not know it poses a risk. Here is what the problem looks like and how to solve it.

Attackers manipulate AI prompts to exfiltrate data from your company — using legal tools to do it. How does prompt injection work and why is traditional DLP not enough?