True CyberChampion
Upcoming: Monday, October 26, 2026

Trustworthy AIin Practice

Role-Based Training for Ethical, Explainable AI That Reduces Enterprise Risk. AI never shows an error code. When it is wrong, it sounds exactly as sure as when it is right. This course trains every person who uses, approves, buys, or oversees AI to catch the quiet failures, explain the results, and prove AI was used responsibly.

Reserve a seat See the course outline
Next session
Monday, October 26, 2026
Formats
2 hours, 4 hours, or a full day
Delivery
On site, live virtual, or hybrid
Method
Capability Proof System™

Course description

The biggest AI risk is a quiet one

Know when to trust it.
Know how to prove it.

AI never crashes and never shows an error code. It always gives an answer, and when that answer is wrong it looks exactly as confident as when it is right. The real risk is rarely a dramatic failure. It is a quiet one: a biased eligibility flag, a confident but invented citation, a chatbot that gives a customer the wrong deadline, a summary that leaks personal data. Each one passes straight into a real decision because nobody was trained to stop it.

Trustworthy AI in Practice is not another awareness video. It is role-based AI ethics training that changes how people work on Monday morning. Each participant practices on realistic tasks from their own job, whether they are an executive, a program manager, an analyst, an auditor, or a contracting officer, and leaves with the habits, checklists, and decision rights to use AI ethically and explain it.

Participants learn the principles of trustworthy AI, apply the four principles of explainable AI, run quick bias and privacy checks, and document AI decisions that can be defended. Every participant is measured before and after, so your organization receives evidence of real skill, not just attendance. The course is designed to meet annual AI ethics and trustworthy AI training requirements.

Objectives and outcomes

What the course sets out to do, and what your people can do after

Every outcome is practiced on realistic tasks from each participant's own job, not demonstrated on slides.

Course objectives

What this course is built to achieve for your organization

  • Give every role one shared picture of trustworthy AI: seven principles, the four principles of explainable AI, and the four risks untrustworthy AI creates.
  • Turn principles into daily habits through hands-on practice on each participant's own work.
  • Set clear rules for what AI may do, with which data, role by role.
  • Make every AI-assisted decision explainable and auditable with a simple record.
  • Prove skill change with before-and-after measures, not attendance.
  • Meet annual AI ethics training requirements with records leaders and auditors can rely on.

Learning outcomes

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

  1. Explain why AI fails differently from traditional software, and name the four risk categories it creates.
  2. Define ethical, responsible, and trustworthy AI, and apply the seven principles to their own work.
  3. Test any AI output against the four principles of explainable AI.
  4. Run the five-question TRUST Check before using or acting on AI output.
  5. Recognize bias in data, models, and use, and perform a fairness check suited to their role.
  6. Protect personal and sensitive data, using approved tools and the data rules for their role.
  7. Document an AI-assisted decision so it is transparent, traceable, and ready for audit.
  8. Escalate concerns at the right point in the AI lifecycle, and commit to specific changes tracked for 30 days.

Course outline

Ten modules, one practical path

The same modules scale from a 2-hour annual requirement to a full-day lab. Pick a format to see the time for each module. Open a module to see what participants do.

  1. 0Opening and baselineMeasure where each person starts
    Each participant answers a short, real-looking scenario for their role before any teaching: What would you do with this output? What could go wrong? What would you record?
    • Baseline scenario, scored later with the same rubric
    • The gap between using AI and knowing the rules
  2. 1AI fails quietlyWhy AI needs a different kind of trust
    Loud failure versus quiet failure, and the three properties that make AI different: it always answers, its answers vary, and its reasoning is hard to see.
    • The four risks: strategy and reputation, cyber and privacy, legal and regulatory, operations
    • Exercise: spot the quiet failure in three realistic outputs
  3. 2The seven principles and their tradeoffsEthical, responsible, and trustworthy AI
    Seven principles, each reduced to one question people remember: Purposeful, Fair, Reliable, Transparent and explainable, Accountable, Privacy, Safe and secure.
    • Tradeoffs: accuracy versus explainability, fairness versus privacy, speed versus reliability
    • Exercise: principle sort with incident cards
  4. 3Explainable AI in practiceThe four-principle test
    Explanation, meaningful, explanation accuracy, and knowledge limits. A right answer with the wrong reasons will fail as soon as conditions change.
    • Explanations matched to the audience: builder, manager, or affected person
    • Lab: test an AI output from your own role, then write the plain-language explanation
  5. 4Fairness and bias checksWhere bias enters, and who checks for it
    Bias enters through data, through stand-ins that track protected characteristics, and through use on the wrong population.
    • The proxy problem, with a real-world example
    • Lab: read a fairness table and decide to continue, add review, or pause
  6. 5Privacy, security, and data governanceThe first decision is the data
    Minimize, redact, then use. AI-specific threats including prompt injection, data leakage, and over-permissioned agents.
    • Lab: redact before you prompt, then choose the right tool
    • Builders and security staff: review an agent's permissions
  7. 6Transparency, documentation, and the audit trailIf it is not documented, it did not happen
    A short AI decision record for everyday use, and the system-level records leaders and auditors need.
    • Disclosure: when people should know AI was involved
    • Exercise: write a record, then test whether an auditor could reconstruct the decision
  8. 7Governance, compliance, and the lifecycleYour checkpoints from design to retirement
    Four lifecycle phases, the checkpoints that belong to each role, and three clear triggers for escalation.
    • Keeping up with changing rules without memorizing them
    • Exercise: follow one AI use case through the lifecycle
  9. 8Role lab: the Delegation Grid™ in actionPractice on your own work
    Participants sit in role tracks and work a complete scenario from their own job, running the TRUST Check at every step.
    • Place three real tasks on the Delegation Grid: delegate, co-pilot, or human only
    • Turn the result into a commitment for your next real use of AI
  10. 9Prove and commitMeasure the change, then make it stick
    The baseline scenario again, scored the same way, so each person sees their own change.
    • Knowledge check and annual attestation
    • Launch Ledger: two or three commitments, with a 30-day manager check-in

The 4-hour format includes a 10-minute break. The full day runs 6.5 hours of instruction, excluding lunch and breaks. The executive course uses the same content, re-sequenced for leaders.

Role-based by design

Same principles. A different practice for every role.

Seven role tracks. Pick one to see what it practices and what it takes back to work.

Practices
Leaves with

Formats

Choose the format that fits your team

On site, live virtual, or hybrid. Every format includes the baseline, a role lab, and the proof step, so every format produces evidence.

2 hours

Core annual training

For all staff who use AI tools. Designed to meet annual AI ethics and trustworthy AI training requirements.

Choose this format
4 hours

Executive course

For senior leaders and boards. What to approve, what to ask, what to own, and what proof to expect, with a live tabletop exercise.

Choose this format
4 hours

Practitioner workshop

For people who approve, build, buy, secure, or audit AI. Deeper labs on bias, explainability, and records.

Choose this format
Full day

Trustworthy AI Proof Lab

For AI governance teams and new AI programs. Two scenarios per track and a confirmed role-by-role rule set.

Choose this format

What your organization receives

Training records that prove something

Ready for your learning management system, your auditors, and your annual training report.

  • Completion rosterBy person, role, format, and hours
  • Signed attestationsStaff and leadership versions
  • Before-and-after scoresSkill change, by role
  • 30-day commitmentsFor manager follow-up
  • Leadership summaryRisks staff raised and the policy fixes they point to
Dr. Beza Lefebo, CEO and Chief AI Officer

Lead instructor

Dr. Beza Lefebo

CEO and Chief AI Officer, True CyberChampion. Doctor of Engineering from The George Washington University, where his research built explainable AI to detect cyberattacks on the U.S. smart grid. CISSP, CISM, and CDPSE. Dr. Beza brings decades of experience from federal, big tech, international assignments, and defense contracting, and has researched, written, and consulted extensively on trustworthy AI.

Built on public guidance that government buyers already trust:

  • NIST AI Risk Management Framework
  • NIST Four Principles of Explainable AI
  • HHS Trustworthy AI Playbook
  • OECD AI Principles

Questions

Questions buyers ask

Anything else, just ask when you register.

What happens on October 26, 2026?

That is the next scheduled session. Reserve a seat or a group booking below, and we will send the time, location or joining link, and pre-work.

Does this meet our annual AI ethics training requirement?

It is designed to. Every format produces a completion record and signed attestation. Confirm the hours and topics with your training office; we tailor the content to the policies that apply to you.

Can you run it on our own date, with our own AI tools and policies?

Yes. For group bookings we rebuild the scenarios from your organization's tasks, approved tools, and rules, with sensitive details removed, and schedule a date that suits you.

Who is it for?

Federal, state, and local government, government contractors, and regulated industries. Content is built from real public-sector and enterprise work.

How large can a group be?

Labs work best with 15 to 35 participants per facilitator. The executive course scales to larger rooms with table facilitators.

Is it virtual or in person?

Either, or hybrid. Virtual sessions use breakout rooms by role so every participant still practices.

Monday, October 26, 2026

Register

Reserve a seat, or bring the course to your team

Tell us your audience, format, and timing. We confirm session details on registration, and send group quotes within five business days.

Government buyers
CAGE 9DE87 · UEI UEQDZSKKRFGC65

We typically respond within one business day.