Demystifying AI · Stage 6

AI safety, security and rules

โ† Demystifying AIStage 6: Using AI wellNot yet checked by our team

AI brings new risks alongside the usual ones. Prompt injection hides instructions in text the model reads, such as a web page or an email, to make it ignore its real instructions. Models can leak sensitive data they were given, and poisoned training data can plant bad behaviour. Defences: give an AI only the access it needs, keep secrets and personal data out of prompts unless they're protected, filter what goes in and out, log what it does, and have people approve risky actions. Rules are arriving too, such as the EU AI Act and the NIST AI Risk Management Framework, so organisations keep records of their data, models and decisions (AI governance).

Why it matters: Every AI system that touches real data or real systems needs these defences.

Builds on: Prompting well

Learn it, then prove it

Sign in (free) to take this step's mastery check and save your journey.


Contact Us

Questions or feedback about anything? We'd love to hear from you.

You can reach us at: support@seriousmindware.com