School: Minister Destiny Testimony Schools Track: Adult Class: Certificate Topics — Year 1 (Foundations) · Home: Advocacy Certificate behind this topic: "Using Generative AI in Public Relations" — LinkedIn Learning, earned by Minister Destiny Marie Morgan (Jun 06, 2025) Founded and presented by: Muslim Ordained Minister Destiny Marie Morgan
This lesson is educational information from Minister Destiny Testimony Schools. It teaches the public body of knowledge in this subject — the same ground Minister Destiny studied for her certificate — in this school's own words. It is not the original course, it does not award the original credential, and it is not legal or medical advice. If you are in danger right now, call 911.
Before you start
- This school never asks your name. Your progress is saved only in your own browser on this device — the school collects and keeps nothing. If you pass this unit's quiz (80% or higher, retakes allowed), you earn a downloadable certificate with a blank line — you write your own name on it after you save it. Quiz answers are not shown; if you miss a question, re-read the lesson and try again.
- One rule runs through this unit, non-negotiably: verify everything the machine offers, disclose assistance where the audience deserves to know, and never publish a fabricated quote, statistic, or testimonial. The guardrails are not an appendix to this topic; they are the topic.
- Minister Destiny studied this subject as AI tools entered advocacy work. The working rule here: the tool may draft, but a named human answers for every published word.
The lesson
Part 1. What Generative AI Does Well — and Badly — in PR
Generative AI tools produce text, images, and audio from prompts, predicting plausible output from training patterns. In PR they are genuinely good at first drafts and alternative phrasings, brainstorming angles for a human to judge, summarising documents you supply, adapting finished text to a new length or reading level, and routine variations where the facts are fixed and known. Used there, they return hours to the work only humans can do: relationships, judgement, and verification.
They are reliably bad at what PR cannot afford to get wrong. They fabricate: invented quotes, sourceless statistics, events that never happened, citations to articles that do not exist — in the same confident tone as true material; practitioners call this hallucination. They are stale: training ends at a cutoff, so current facts are outside their knowledge unless checked against live sources yourself. They flatten voice toward their training average, and mirror bias in their data — in advocacy copy, stereotypes can arrive dressed as neutral drafts. The tool is a fast, tireless, unverified junior. Staff it like one.
Part 2. Workflows — Drafting and Brainstorming With a Human in Charge
Safe AI workflows share a shape: the human supplies verified material and judgement; the machine supplies speed and variation; the human verifies and decides before anything leaves. In drafting, brief from facts: give the tool only confirmed facts, ask for structure or a first draft built from those, then rewrite in your voice with sources beside you. In brainstorming, ask for quantity — twenty headline angles, ten likely objections — and treat output as a list to judge, not a decision: the value is the handful you would not have thought of, selected by you.
Two habits keep the shape safe. Keep verification separate from editing: editing makes text smoother, which quietly smooths its errors too — check facts first (Part 3), polish second. And keep prompts and outputs for significant pieces: that saved record is your accountability trail if a claim is ever questioned. Crisis work tightens everything further: no AI-drafted statement goes out without line-by-line human verification against confirmed sources, because a fluent error in a crisis statement is the most expensive sentence in PR.
Part 3. The Checking Routine — Fact-Checking AI Output
Every factual claim in AI output is unverified until you check it — especially the familiar-sounding ones, since familiarity is how fabrications slip through. The routine has five steps. (1) Extract the claims: every checkable assertion — names, dates, figures, quotes, events. (2) Go to primary sources: the records, the official publication, the data itself — never the AI's offered citation unopened, since invented references are a signature failure. (3) Check quotes word by word against a recording, transcript, or the speaker: a quote never said is a fabrication even if it captures the "gist." (4) Check numbers arithmetically: percentages against their base, totals against their parts, dates against the calendar. (5) Cut what could not be verified: an unchecked claim is not softened with "reportedly" and shipped; it leaves the draft.
Statistics and testimonials get special strictness, because readers weight them heavily. An AI-proposed statistic is traced to a real, named, current source you have opened — cited with its actual finding — or it does not run. An AI-proposed testimonial never runs, in any form: testimonials are evidence of a real person's real experience, and generating one is fabrication by definition. Build the routine into a checklist that travels with the draft, signed off by name — the named-human rule, made mechanical.
Part 4. Disclosure — Telling the Audience What Deserves Telling
Disclosure means letting the audience know, where it could matter to their judgement, that AI assisted what they are reading, hearing, or seeing. PRSA guidance treats transparency as the default: disclose when AI use is material to how content should be weighed. Routine assistance — spelling, formatting, a first draft heavily rewritten and fully verified by the named human publishing it — sits at the light end. The heavy end demands explicit, in-place disclosure: generated or altered images and audio, synthetic voices, content published substantially as generated, and any chatbot a person might mistake for a human.
Three disclosure rules have no calibration. Never present AI output as a human's words — a generated "quote," first-person account, or letter "from a resident" is deception regardless of small print. Never simulate public support — generated comments, reviews, or grassroots voices are astroturfing, a settled ethical violation long before AI made it cheap. And when in doubt, disclose: an unnecessary disclosure costs a sentence; a discovered concealment costs the audience's trust in everything else you publish. Plain disclosure is also an asset: audiences reward organisations that state their practice before being asked.
Part 5. Confidentiality — Never Paste What Isn't Yours to Share
Many AI tools — especially free, consumer versions — may store what you enter, use it to improve their systems, or expose it through account weaknesses. The rule follows directly: never paste private data into a tool whose handling you have not verified. That covers other people's personal information, unpublished organisational material (draft statements, internal figures, legal advice, embargoed announcements), and anything received in confidence. A breach by copy-paste is still a breach, and "the AI tool" is not a defence anyone accepts.
Workable habits make the rule livable. Redact before you prompt: replace names and identifying details with placeholders, restoring them only in your own document. Prefer approved tools: name which tools may be used, at which tiers, with what data — and without an approved option, treat every tool as public. Mind the by-products: uploaded documents, connected drives, and extensions widen what a tool can see. In advocacy work, where people hand you the worst facts of their lives in trust, confidentiality is the substance of that trust; AI convenience never outranks it.
Part 6. Images, Organisational Policy, and the Small-Team Version
Generated images and audio carry their own limits. An image is an illustration, never evidence: never present it as a photograph of a real event, place, or person; label it where it could be mistaken for documentation; never imitate an identifiable person doing or saying anything. Check visuals as carefully as text — wrong signage, invented details in a real place, stereotyped depictions of your own communities. Audio is stricter: a synthetic voice of a real person requires explicit consent, and many uses — a leader "saying" words never recorded — should simply be refused.
Finally, write it down. An organisational AI policy at small scale fits on one page: approved tools and uses; the never-paste list (Part 5); the verification checklist and its sign-off (Part 3); the disclosure practice with its three uncalibrated rules (Part 4); the image and audio lines (this Part); and who answers for published output — a name, because accountability belonging to "the team" belongs to no one when someone asks. Review it when tools change. Its point is that everyone publishing under the organisation's name — including the tired volunteer at midnight — can find the guardrails in one place, in plain words, before they need them.
Words to know — Using Generative AI in Public Relations
- Generative AI — tools producing text, images, or audio from prompts by predicting plausible output.
- Hallucination — AI fabricating material — quotes, statistics, events, citations — confidently, with no signal it is invented.
- Verification pass — the separate step checking every claim against primary sources first.
- Primary source — the original record itself — data, transcript, official document — not a summary.
- Disclosure — telling the audience, in place, that AI materially assisted the content.
- Astroturfing — simulating grassroots support with generated comments or reviews — an ethical violation AI makes cheap, not acceptable.
- Redaction — replacing names and identifying details with placeholders.
- Synthetic media — AI-generated or altered images, audio, or video; illustration, never evidence.
- Training cutoff — where a model's knowledge stops; current facts need checked sources.
- Named-human rule — a specific person answers for every published word, whatever drafted it.
Myths vs facts — Using Generative AI in Public Relations
- Myth: If the AI states it confidently, it is probably accurate. Fact: Confidence is the tool's tone, not its evidence. Fabricated quotes, statistics, and citations arrive in exactly the same voice as true material — which is why every claim is unverified until you check it.
- Myth: A quote that captures what someone "would say" is fair to publish. Fact: A quote the person never said is a fabrication, gist or not. Quotes are checked word by word against a recording, transcript, or the speaker — or cut.
- Myth: Disclosure makes you look less professional. Fact: Plain, in-place disclosure where AI use is material is the professional standard, and audiences reward organisations that state their practice before being asked.
- Myth: Pasting case details into a free AI tool is private, like typing in a document. Fact: Consumer tools may store and reuse inputs. Never paste private data into a tool whose handling you have not verified; redact first, and prefer approved tools.
- Myth: A generated image is just a quicker photograph. Fact: A generated image is an illustration, never evidence. It must not depict real events or identifiable people as if documenting them, and needs labelling wherever it could be mistaken for a photograph.
Think it through — Using Generative AI in Public Relations
A volunteer drafts your campaign update with an AI tool and sends you a polished text: it quotes your director saying "no family in this borough should face eviction alone" (she believes the sentiment; she has never said the sentence), cites "a 2025 study showing 68% of tenants win when advised" (no source named), and includes a warm testimonial from "Sarah, a grateful client" (the tool composed it). The volunteer also pasted three real client case summaries into the tool to "give it the flavour." It reads beautifully and the deadline is tonight.
Walk it through: (1) Workflows (Part 2): the draft skipped the shape — no verified material supplied, machine variation treated as decision. You, the named human, now own the checking, in separate verification and editing passes. (2) Checking routine (Part 3): the quote fails word-by-word checking and is cut, or taken to the director to actually say on the record. The statistic has no openable primary source and is cut, not softened. The testimonial is fabrication by definition and is cut entirely. (3) Confidentiality (Part 5): the pasted case summaries are the most serious breach here — establish what tool received them, follow your incident steps, and retrain the volunteer kindly but completely. (4) Disclosure and policy (Parts 4, 6): the published piece carries your stated AI practice, and the checklist gains a named sign-off before any AI-assisted draft ships again.
Try this — Using Generative AI in Public Relations
- ☐ Write your five-step checking routine as a checklist that travels with drafts
- ☐ Generate a practice draft, extract every checkable claim, then verify or cut each against a primary source
- ☐ Draft your disclosure practice in three sentences: your general statement, what gets labelled in place, and your never-rules
- ☐ Write your never-paste list for your actual work, and practise redacting one document with placeholders
- ☐ Write a one-page AI policy: approved tools, never-paste list, verification sign-off, disclosure practice, image and audio lines, and the accountable person
Resources to keep — Using Generative AI in Public Relations
- Promise & Pitfalls: The Ethical Use of AI for Public Relations Practitioners — Public Relations Society of America (prsa.org) — the profession's own published AI ethics guidance: transparency, human oversight, verification, and confidentiality.
- Code of Ethics — Public Relations Society of America (prsa.org) — the underlying duties — honesty, disclosure, safeguarding confidences — that AI practice is built on.
- Institute for Public Relations (instituteforpr.org) — free research on AI, trust, and misinformation.
- Federal Trade Commission (ftc.gov) — Endorsement Guides on deceptive claims, fabricated testimonials, and disclosure.
What you earned / what's next — Using Generative AI in Public Relations
✅ Using Generative AI in Public Relations complete (quiz score 80% or higher): You can use generative AI for drafting and brainstorming inside a human-in-charge workflow, run the five-step checking routine on any output, disclose assistance where it matters, protect confidential material absolutely, govern generated images and audio, and write a one-page AI policy for a small team.
Your certificate
Minister Destiny Testimony Schools Certificate of Completion — Using Generative AI in Public Relations
This certificate is presented by Muslim Ordained Minister Destiny Marie Morgan
Awarded to: ______________________________ (write your name here after you download — your name is never collected or stored by this school)
Date completed: ______________
You have learned to use a fast tool without surrendering the slow virtues — verification, disclosure, and other people's privacy. The machine may draft, but your name answers. Answer proudly, and check first. — Minister Destiny Marie Morgan
This is a certificate of educational completion from Minister Destiny Testimony Schools. It is not a state license, a government certification, or the original course credential.
What's next: Return to the Certificate Topics for your next Year 1 Foundations topic.
Sources
- Promise & Pitfalls: The Ethical Use of AI for Public Relations Practitioners — Public Relations Society of America, Board of Ethics and Professional Standards (prsa.org) — used for transparency, disclosure, human oversight, verification, and confidentiality guidance.
- Code of Ethics — Public Relations Society of America (prsa.org) — used for honesty, disclosure of information, and safeguarding confidences as the duties beneath AI practice.
- Research on AI, trust, and misinformation in communication — Institute for Public Relations (instituteforpr.org) — used for audience trust and disclosure findings at orientation level.
- Endorsement Guides — Federal Trade Commission (ftc.gov) — used for the rule against fabricated testimonials and undisclosed material connections.
Your Progress — On Your Device Only — Using Generative AI in Public Relations
- Learning outcomes: On completing this unit, the student can:
- describe what generative AI does well and badly in PR, including hallucination and training cutoffs
- run a drafting and brainstorming workflow in which a human supplies verified material, verifies, and decides
- apply the five-step checking routine, including word-by-word quote checks and trace-or-cut statistics
- apply disclosure calibration and the three rules with no calibration: no AI words as a human's, no simulated support, disclose when in doubt
- enforce confidentiality: never paste unverified private data, redact first, prefer approved tools
- govern generated images and audio and write a one-page organisational AI policy with a named accountable human
- Study time: Approximately 2 hours — an estimate covering reading, activities, and the test. Students move at their own pace.
- Assessment: A 10-question test drawn only from this unit's lesson. 80% or higher passes and earns the unit certificate. Answers are not shown; your score is saved only in your own browser on this device, and retakes are allowed.
- Level: Certificate Topics — Year 1 (Foundations) · Adult school · working rank on completion path: Freshman → Sophomore.
- Version: Version 1.0 — October 2026.
Minister Destiny Testimony Schools — Adult. Educational information only. Help lines: 911 for emergencies · 988 Suicide & Crisis Lifeline (call/text 988).