Wise AI Ambassador · Set 17 of 20
Protection
Cards for the fourth degree of the Wise AI Ambassador
Tell technical attacks on systems from cognitive attacks on people, and build a shield against each.
A card for each unit of the degree, with its idea, its exercise and its takeaway, and the gate's card at the end.
style7 cardsschedule1 min read



Protection
Tell technical attacks on systems from cognitive attacks on people, and build a shield against each.
A model is a new attack surface.
Map where data enters an AI system you use, and where it leaves.
Every input to a model is a door that needs a guard.
A model is a new attack surface.
Map where data enters an AI system you use, and where it leaves.
Every input to a model is a door that needs a guard.
tocCards
Protection
Tell technical attacks on systems from cognitive attacks on people, and build a shield against each.
Copyright
Protection
Metn · Wise AI Ambassador
First Edition – 2026
Author: Prof. Dr. Mohamed Fawzi Elgendi
© 2026 Prof. Dr. Mohamed Fawzi Elgendi. All rights reserved.
This work is available under the open Fawzooz Waqf Licence. Use, development and circulation for lawful purposes are permitted, with attribution to the original creator required.
Published by: FAWZOOZ
London, United Kingdom
Company Registration No: 17511660
www.fawzooz.ai
Tell technical attacks on systems from cognitive attacks on people, and build a shield against each.
Tell technical attacks on systems from cognitive attacks on people, and build a shield against each.
A card for each unit of the degree, with its idea, its exercise and its takeaway, and the gate's card at the end.
A model is a new attack surface.
Map where data enters an AI system you use, and where it leaves.
Every input to a model is a door that needs a guard.
Know the attack to know its shield.
Choose one attack, and write a realistic scenario for it in your organisation.
What you cannot name, you do not see when it happens.
Attack your system before someone else does.
With permission and in a test environment, try to make an assistant break its instructions, and record what happened.
A flaw you find yourself costs less than one someone else finds.
The human mind is the easiest target.
Look back at the last urgent message you received, and pick out the pressure it put on your feelings.
Slow down: haste is the attacker's first weapon.
A voice and a face are no longer proof.
Agree a verification word, or a second channel for confirming, with your team or your family.
Call back yourself; don't rely on the one who called you.
Prepare for encryption that may break tomorrow.
Ask your technical team: which of our data must stay secret for ten years?
What must stay secret for a long time must be protected from today.
An adversarial testing report
Test at least one technical attack and one cognitive attack, in a safe, authorised environment.
When you record what you found and its impact, and propose a shield for each.

