The risks the training covers
Data leaking into AI tools
Employees paste sensitive information, code or customer data into ChatGPT/Copilot - and it leaves the organization's control.
Prompt injection
An attacker embeds malicious instructions in text, a document or a website that the AI reads, and makes it act against the organization.
Shadow AI
Using unapproved AI tools, without control, policy or an understanding of the privacy risks.
Wrong or invented output
Blind reliance on wrong AI answers ("hallucinations") in business decisions or in code.
Two tracks
For employees
- What is allowed and what is forbidden to paste into AI tools
- Recognizing prompt injection and AI-based social engineering
- A policy for the correct use of Copilot / ChatGPT / Gemini
- Protecting personal information and privacy (Amendment 13) when using AI
For developers
- Secure AI SDLC - secure development of AI capabilities
- OWASP LLM Top 10 - the main risks in LLM applications
- Prompt injection, system prompt leakage and data poisoning
- Controls: sandboxing, least privilege, input/output filtering
Run a phishing simulation for your employees yourself - in 15 minutes* · from ₪999
* One allowlisting action by your IT team is required, and launch is subject to a security review of up to one business day.