Module 1 · Course Overview & Syllabus

Course Overview & Syllabus

Who this course is for, what it covers, and the one line every module keeps returning to: AI can check your work, it cannot write it.

Who This Course Is For

This course is built for three roles inside a municipal law enforcement agency, and it teaches each of them differently:

  • Patrol Officers — who need a bright line for what they can and cannot let an AI tool touch on their own reports
  • Record Managers — who review, certify, and disclose reports, and who are the last checkpoint before a document becomes part of the evidentiary record
  • Command Staff — who set agency policy, approve procurement, and answer to both the city and the county prosecutor’s office when something goes wrong

If you supervise, review, or personally author incident narratives, this course is written for your job specifically — not as a generic “AI in policing” survey.


Why This Course Exists Right Now

Two things are true at the same time in King County, and this course exists in the gap between them:

  1. The King County Prosecuting Attorney’s Office (KCPAO), under Chief Deputy Prosecutor Daniel J. Clark, has told law enforcement partners in writing that it will not accept incident narratives produced with the assistance of generative AI tools like Axon Draft One — full detail in Module 2.
  2. The City of Seattle’s 2025–2026 AI Plan actively directs city departments toward AI adoption — specifically workforce upskilling and university partnerships — for tasks like data integration, privacy compliance, and municipal policy work. Full detail in Module 4.

Those two positions are not in conflict. They describe two different jobs AI can do:

AI-as-AuthorAI-as-Auditor
What it doesGenerates new narrative prose from audio/videoChecks already-written, human-authored text against a policy or evidentiary checklist
KCPAO positionRejected — the reports “will not be accepted”Not addressed by the directive — no prose is generated
Where it fits this courseExplicitly out of scope, and actively discouragedThe entire operational model taught in Module 5

This course teaches you to operate entirely on the right side of that table.


Course Objectives

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

  1. Explain, in legal terms, why AI-authored narrative text creates prosecution risk — specifically how it can create Brady and Giglio exposure for the officer who signs the report.
  2. State the KCPAO’s directive accurately and know what specifically it prohibits (and does not prohibit).
  3. Cite the actual research on whether AI report-writing tools save time — and correct vendor claims that don’t hold up.
  4. Identify which digital tools are safe to use (transcription, structured-checklist QA) versus which cross into prohibited territory (narrative generation, pasting case data into consumer chatbots).
  5. Walk through a compliant, AI-as-Auditor workflow step by step, using tools that check an officer’s own words rather than replacing them.
  6. Recognize gray-area scenarios — short-staffed shifts, ambiguous audio, time pressure — and make the compliant call under realistic operational pressure.

The Line We Keep Returning To

Every module in this course will return to one sentence. Learn it now:

AI may check whether you said what needs to be said. AI may never say it for you.

Transcription is fine. Spell-check is fine. A checklist tool that asks “did your narrative document the specific, articulable facts supporting reasonable suspicion?” is fine — it doesn’t write a word of your report. A tool that drafts the paragraph describing the stop, the search, or the suspect’s statements is not fine, regardless of how good the vendor’s demo looks.

⚠️ Common Pitfall: This Is Not a Ban on AI

Officers sometimes hear “the county won’t accept AI reports” and conclude the department is anti-technology. That’s the wrong takeaway, and it’s not what either the KCPAO directive or the Seattle AI Plan says.

The KCPAO objects to a specific failure mode: AI systems generating prose about events, statements, and observations that the officer did not personally verify word-for-word before signing. The City of Seattle, in the same policy window, is actively funding AI upskilling for other work.

The skill this course builds is discrimination, not avoidance — knowing exactly which AI use case you’re looking at before you touch the tool.

💻 Code Lab: Map Your Own Agency's Tools

Before Module 2, take an inventory. For every digital tool your agency currently uses that touches incident reports, evidence, or case data, answer:

  1. Does this tool generate new narrative text, or does it check text that a human already wrote?
  2. If it generates text, whose words end up in the signed report — yours, or the vendor’s model’s?
  3. Has your agency’s records division or command staff formally reviewed this tool against the KCPAO directive?

Keep your answers. Module 5 will ask you to apply the AI-as-Auditor framework to this same list.


Up Next

Module 2 goes straight to the legal mechanics: why a single AI-invented detail in a signed report can trigger Brady and Giglio obligations that follow an officer for the rest of their career — testifying in this case and every case after it.