AI at Work: Safe, Secure & Compliant

• One day, instructor-led, with short practical labs throughout.
• Written for every role, not just technical staff - no programming experience is needed.
• Covers public, enterprise and custom AI assistants, and the difference that matters for your data.
• Addresses Shadow AI, automation bias and the habit of critically evaluating AI-generated output.

• No technical background is required.
• General familiarity with everyday business tools and a browser is enough.
• No prior experience with ChatGPT, Gemini, Claude or Copilot is assumed.

By the end of this course, participants will be able to:

• explain in plain language how generative AI produces answers, and why it can be confidently wrong
• recognize AI-enhanced phishing, spear phishing, deepfake video and voice fraud
• describe how AI lowers the barrier for less skilled attackers and how malware evades detection
• identify which categories of company information must never be entered into a public AI tool
• explain the risks of data retention and model training, and the risks of Shadow AI
• distinguish public from enterprise AI services and apply the safe-use rules for their own role
• recognize privacy and legal exposure: personal data, GDPR, DPIAs, the EU AI Act and copyright
• spot automation bias, authority bias and overreliance in their own use of AI
• recognize prompt injection and the disclosure risk it creates
• configure a simple custom assistant, and judge when one adds value or adds risk

• Module 1. Introduction
  • What AI is, and why it is a tool rather than magic.
  • How generative AI works and how it differs from traditional software.
  • Why AI has no true understanding: pattern recognition, hallucinations and misinformation.
  • Lab: prompting best practices.
• Module 2. AI and Cybercrime: What Has Changed
  • AI-enhanced phishing and spear phishing built from public social media.
  • Deepfake video and voice fraud, including deepfake injection during live calls.
  • AI-assisted malware, polymorphic code and the lowered barrier for unskilled attackers.
  • Lab: prompt guardrails.
• Module 3. Using AI Safely at Work
  • Information that must never be entered into public AI tools, and why.
  • Data retention, model training and vendor dependency.
  • Shadow AI: the risks, and how managers recognize it.
  • Public versus enterprise AI solutions; a high-level view of GDPR, HIPAA and similar obligations.
• Module 4. AI, Privacy and Legal Risks
  • Personal and sensitive data in an AI context, and privacy risk at the point of entry.
  • Privacy and AI impact assessments, including DPIAs.
  • The EU AI Act's risk-based approach, and why U.S. requirements vary by federal, state and sector.
  • Copyright, intellectual property, liability and contract exposure; bias and discrimination.
• Module 5. The Human Factor: The Greatest Vulnerability
  • Automation bias and authority bias, and recognizing overreliance on AI output.
  • Checking sources and verifying AI output logically.
  • Prompt injection and the disclosure of confidential information.
  • How AI errors spread quickly and at scale - the flywheel effect of automated mistakes.
• Module 6. From Risk to Practice: Working Safely with Custom AI Assistants (time permitting)
  • What a custom AI assistant is and when it adds value.
  • Platform differences across ChatGPT, Gemini, Claude and Copilot.
  • What information can safely be used, and when assistants create privacy and security risk.
  • Labs: create a Custom GPT; create a Gemini Gem.