Certified AI Executive: Chief Information Security Officer

What You Will Be Able to Do

• Explain what AI can and cannot do, test a team's or vendor's claims, and count the true cost of using it
• Choose where AI creates value, build a business case on measured baselines, and move pilots to scale or stop them
• Set accountability, risk controls and compliance for AI using recognized frameworks and regulation
• Lead the organization through AI: operating model, skills, workforce commitments, change and adoption
• Explain how AI systems are attacked and set priorities using the OWASP Top 10 lists and MITRE ATLAS
• Build security into AI systems, their supply chain and agent permissions, and fit AI into an existing security program
• Counter AI-enabled fraud and phishing, use AI in the security operations center safely, and measure detection and response honestly
• Lead the response to AI incidents, support breach notification and disclosure, and report risk, budget and skills to the board

Modules

1. Part 1: Executive Core

   Part 1 is the same in every executive certification.

2. 1AI Essentials for Executives

   Leaders do not need to build AI, but they do need to judge it. This module explains the main kinds of AI system, what models can and cannot do, why data sets the limits, how AI costs behave, and the vocabulary and tests a leader needs to question a team or a vendor.

   • Rule-based systems, machine learning, generative AI and AI agents
   • What models do and where they fail: errors, bias, drift and explainability
   • Data, economics and the vocabulary for questioning teams and vendors
   • Spotting hype and testing vendor claims on your own data

   Exercise. List the AI tools and features already in use in your area, including those built into software you buy. For each, note whether it is rule-based, machine learning, generative or an agent, what data it uses, and who reviews its outputs before they are acted on.

3. 2AI Strategy, Value and the Business Case

   AI is worth funding only where it changes a measure that matters. This module covers where AI creates value, how to build and rank a portfolio of use cases, when to build, buy or partner, how to write a business case that counts the full cost and tracks realized benefit, and how to move from proof of concept to scale or stop.

   • Where AI creates value and how to measure it
   • Use-case portfolios: value, feasibility and risk; build, buy or partner
   • The business case: baselines, total cost of ownership, payback and ROI
   • Proof of concept, pilot and scale; why pilots stall; stage gates

   Exercise. Take one AI proposal now under discussion in your organization. Write down its baseline measure, list every cost it will carry over three years (including data work, integration, change, monitoring and review effort), name the owner of each benefit, and set the criteria for the next stage gate.

4. 3AI Governance, Risk and Regulation for Leaders

   AI needs clear owners, a view of what is in use and controls that match the risk. This module covers accountability structures, risk tiering, the main frameworks and standards, how AI and privacy law apply in the US, Canada and the EU/UK, and the confidentiality, security, vendor and incident risks that leaders must manage.

   • Accountability, the AI governance committee, RACI and the three lines
   • AI inventory, risk tiering and risk appetite
   • NIST AI RMF, ISO/IEC 42001, and AI and privacy law across jurisdictions
   • Confidentiality, AI security, vendor risk and incident response

   Exercise. List the AI systems and tools used in your area, including AI features inside software you already buy. For each one, name the business owner, assign a risk tier using the questions in this module, and note one control that is missing.

5. 4Leading People and the Organization Through AI

   AI value depends on how people and work are organized around it. This module covers operating models and leadership roles, AI literacy, workforce impact and honest commitments, change management, shadow AI and acceptable use, and how to measure real adoption and lead by example.

   • Operating models and AI leadership roles
   • AI literacy and workforce impact: tasks, redesign and reskilling
   • Leading change: Kotter's eight steps, involvement and resistance
   • Shadow AI, acceptable use, real adoption and communication

   Exercise. Pick one team that has access to an AI tool. Find out how many people used it in the last week, which tasks they used it for and what stops the others, then agree two changes with the team and measure use again a month later.

6. Part 2: Chief Information Security Officer

   Written for chief information security officers.

7. 5Threats to AI Systems

   AI systems can be attacked in ways that ordinary applications cannot. This module looks at those threats from the CISO's seat: why AI needs its own threat model, how direct and indirect prompt injection work, how data and models are poisoned or stolen, why model output and agent actions must be treated as untrusted, and how to use the OWASP Top 10 lists and MITRE ATLAS to set priorities.

   • Why AI systems need their own threat model
   • Prompt injection, direct and indirect
   • Poisoning, model theft, extraction and hidden context exposure
   • Improper output handling, excessive agency and using the reference lists

   Exercise. Pick the three AI systems in your inventory that read the most outside content. For each, write down what untrusted content it reads, what sensitive data it can reach and what actions it can take, and name the one control that would most limit the damage if an injected instruction succeeded.

8. 6Securing AI in the Enterprise

   Most AI security failures are decided before launch, by what a system is allowed to reach and how it was tested. This module covers secure-by-design practice for AI, the AI supply chain of models, data sets and plug-ins, data leakage through prompts and retrieval, identity and permissions for agents, testing and red-teaming, and how to fit AI into an existing program built on NIST CSF 2.0 and ISO/IEC 27001.

   • Secure by design for AI and the AI supply chain
   • Data leakage through prompts, retrieval and logs
   • Identity and permissions for AI agents
   • Testing, red-teaming and mapping to NIST CSF 2.0 and ISO/IEC 27001

   Exercise. Take one AI system that is about to launch and list every identity it uses, every data source it can read and every action it can take. Mark each one that the task does not strictly need, and agree with the owner on which to remove before launch.

9. 7AI Used by Attackers and Defenders

   AI makes familiar attacks cheaper, faster and more convincing, and it gives defenders new tools with their own risks. This module covers how AI changes the attacker's economics, deepfake and voice-clone fraud, AI-written phishing, AI in the security operations center, honest measures of detection and response, and how to judge AI security products.

   • How AI changes the attacker's economics
   • Deepfake, voice-clone and AI phishing attacks, and the controls that work
   • AI in the security operations center
   • Measuring detection and response honestly and judging AI security tools

   Exercise. Write down the steps your organization follows today when a senior executive asks by phone or video for an urgent payment or a password reset. Mark each step that relies on recognizing a face or voice, and replace it with a check that would still work if the face and voice were perfect fakes.

10. 8Incidents, Disclosure and the Security Program

    When an AI system is attacked or misused, the CISO has to contain it, establish the facts quickly and support decisions on who must be told and when. This module covers AI incident response, breach notification principles with examples from the US, Canada, the EU and the UK, reporting security and AI risk to the board, and building the budget and skills the program needs.

    • Responding to AI incidents
    • Breach notification principles and reporting clocks
    • Examples from the US, Canada, the EU and the UK
    • Board reporting, budget and skills

    Exercise. Choose one AI system that can reach personal data. Write down, for a breach through that system, which laws and contracts would apply, which clocks would start and when, who decides whether to notify, and which facts your team would need to supply within the first 24 hours.