Curriculum · Information Technology
AI-Assisted Cyber Defense
A five-lesson curriculum unit on the defender’s workflow — find, validate, fix, test, prove — with Human + AI embedded and the human leader accountable at every step. Lesson 1 is live now; the remaining lessons are in development.
Unit overview
Cyber defenders protect the systems people depend on — schools, hospitals, water, power, banks. Attackers now move at machine speed, and defenders are answering with AI-assisted security tools: software that scans code for vulnerabilities, drafts patches, and checks that fixes actually work. OpenAI’s Daybreak program is a public, real-world example — it gives verified defenders access to advanced AI models for authorized security work, from reviewing software and investigating suspicious activity to validating vulnerabilities and testing fixes.
This unit teaches the defender’s workflow through the HumanEdge Academy lens: the human decides; the AI advises. Students learn what each stage of defense requires, where AI genuinely helps, where it can mislead, and why a named human must sign off before anything ships. The unit’s through-line is the Academy’s rule — Human Authority → AI Assistance → Human Verification → Human Decision — practiced as “Direct it. Check it. Question it. Decide.”
Unit learning goals
Across the unit, students will be able to:
- Describe the defender’s five jobs — find issues, validate them, fix them, test the fix, prove the work — and explain the cost of skipping any step.
- Evaluate AI-generated security findings for credibility: distinguish real, reachable vulnerabilities from false positives, and explain why reachability matters.
- Judge an AI-generated patch for safety: what it changes, what it could break, and what evidence would justify deploying it.
- Apply “Direct it. Check it. Question it. Decide.” to AI-assisted security work and name the human accountable for the outcome.
- Weigh deploy / revise / hold decisions on a cost-benefit basis and defend the call in writing with cited evidence.
Lesson sequence
Lesson 1 — The Defender’s Job: Find & Validate
An AI security scanner flags a flaw in the school’s attendance app and writes a patch. Students audit the finding — is it real, is it reachable? — and the patch — is it safe? They run Direct → Check → Question → Decide, complete a PACED grid, and sign the deploy / revise / hold decision as the accountable humans.
Lesson 2 — Reading the Scanner: Triage & Validation
Students work a stack of scanner findings — some real, some false positives, some real-but-unreachable. They rank by risk, write the triage rationale, and defend why a “critical” finding can still be a “do not fix” when an attacker can’t reach it.
Lesson 3 — The Patch: AI-Generated Fixes Under Human Review
Given three AI-drafted patches for one vulnerability, students review each diff line by line: what changed, what assumptions the AI made, what tests are missing. They select one patch, write the review notes a senior engineer would require, and state what they would not sign.
Lesson 4 — Prove It: Testing, Regression & Evidence
Students design the test plan for a shipped fix: does the vulnerability close, does anything else break, and what evidence package proves both? They assemble a mini evidence packet — the artifact a real team would file for tracking and auditing.
Lesson 5 — Capstone: The Sign-Off
Teams receive a full case file — finding, validation notes, patch, test results — with one planted flaw. They brief the class, take questions, and sign (or refuse to sign) the deployment authorization. Reflection: the tools keep changing; the leader’s judgment is the constant.
Human + AI integration — the unit thread
“Direct it. Check it. Question it. Decide.”
- Human Purpose — in every lesson, students set the goal first: systems people depend on must stay safe and working.
- Understand the Task · Match the AI · Assign & Instruct — students give the AI precise, bounded assignments (list findings; draft a patch; propose tests) — never the final call.
- Navigate & Direct · Evaluate — students check AI output against evidence: reachability, correctness, side effects.
- Decide · Guide & Improve · Exercise Accountability — a named human signs every deploy / revise / hold decision. The human decides; the AI advises.
Standards map
| Code | Standard |
|---|---|
| 9.R.1 | English 9 (2024): Evaluate sources for credibility and bias, cite sources, ethical and responsible use of all sources including AI. Students audit AI-generated findings and patches. |
| 9.C.2 | English 9 (2024): Speaking and presentation of ideas with audience-appropriate vocabulary. Capstone briefings. |
| 9.W.1 / 9.W.2 | English 9 (2024): Expository writing — thesis, evidence from multiple authoritative sources, concluding section. Decision briefs. |
| EPF.1 | Economics & Personal Finance (2019): Cost-benefit analysis and the PACED decision-making grid. Deploy / revise / hold. |
| 6302 | CTE — Cybersecurity Fundamentals (Information Technology): identify security issues in hardware, software, and data; ethical behavior; privacy and acceptable-use concepts. |
| WRS | Workplace Readiness Skills: WRS-2 Critical Thinking · WRS-10 Teamwork · WRS-11 Big Picture Thinking · WRS-13 Adaptability · WRS-15 Information Literacy · WRS-21 Reading and Writing. |
National Standards Alignment
| Framework | Alignment |
|---|---|
| Career Cluster | Information Technology (National Career Clusters Framework, Advance CTE) — the find-validate-fix-test-prove unit sits inside this national cluster. |
| Career Ready Practices | CRP-11 Use technology to enhance productivity — AI-assisted defense under human authorization; CRP-9 Model integrity and ethical leadership in security response; CRP-8 Utilize critical thinking — validate every AI-flagged threat |
| AI4K12 Big Ideas | Perception — how detection systems sense threats; Learning — how defensive models learn and mislearn; Societal Impact — authorizing AI-assisted response on systems people depend on |
| CSTA K-12 CS | Networks & the Internet and Impacts of Computing concepts (CSTA K-12 Computer Science Standards). |
National framework names verified September 25, 2026 against the Advance CTE National Career Clusters Framework, the Common Career Technical Core Career Ready Practices, and the AI4K12 Five Big Ideas (AAAI/CSTA). No codes were invented.
Careers & credentials
Roles the unit introduces: security operations (SOC) analyst, penetration tester, security engineer, incident responder. Industry credentials students will hear named: CompTIA Security+ (entry-level) and CISSP (experienced professionals). The unit stresses that credentials certify knowledge — judgment is built by directed practice, which is what these lessons rehearse.
References
- OpenAI. “OpenAI extends cyber access to Ukraine for civilian defense.” openai.com, September 2026. — Daybreak gives verified cyber defenders advanced AI for authorized security work: reviewing software, investigating suspicious activity, validating vulnerabilities, testing fixes.
- Reuters. “OpenAI commits $1 billion to cyberdefense effort amid AI safety scrutiny.” September 3, 2026. — The “Daybreak for Frontline Defenders” initiative: subsidized access to AI cybersecurity tools for organizations protecting essential services.
- Human–AI Leadership Framework (H.U.M.A.N.E.D.G.E.). HumanEdge Academy, 2026. — Human Authority → AI Assistance → Human Verification → Human Decision; “Direct it. Check it. Question it. Decide.”; “The human decides. The AI advises.”
- Virginia Department of Education — Standards of Learning (English 9, 2024; Economics & Personal Finance, 2019), Workplace Readiness Skills, and CTE course frameworks (cteresource.org).
This curriculum references OpenAI’s publicly announced Daybreak program as a real-world example. It is an original HumanEdge Academy work and is not affiliated with, sponsored, or endorsed by OpenAI.