Cyber Security

2. Employees using unauthorized apps/services that create security gaps.
4. Corrupting the training data of an AI system so it learns harmful or incorrect behaviour.
6. Attacks exploiting vulnerabilities that the software creator doesn't know about yet.
7. Networks of compromised devices used for spam, DDoS attacks, etc.
8. Employees or contractors misusing access intentionally or accidentally.
10. Slightly modified inputs (images, text, audio) designed to fool AI models into misclassifying things.
12. Forcing an AI system to ignore instructions and output harmful or unintended content.
13. Using leaked passwords from one website to log into another (because people reuse passwords).
14. Using manipulated audio/video to impersonate someone and deceive victims.
15. Hiding malicious data (scripts, payloads, commands) inside images, audio files, or other media so it passes undetected.

1. Intercepting communication between two parties to steal or modify data.
3. Attacking a vendor or service that a company trusts, and spreading malware through legitimate updates.
5. Attempts to bypass safety filters in AI systems using carefully designed prompts or exploits.
9. Fake emails or messages designed to steal credentials or install malware.
11. Trying to clone a machine learning model by repeatedly querying it and recreating it elsewhere.