HOW TO DEAL WITH AI  MEMBERS VAULT

​

A faceless, practical workplace-and-schools learning path from PHPS. Each lesson is 1525 minutes with discussion, a usable checklist, and a plain-language takeaway. Not legal advice. Proof shelf: The Unseen Experiment.

​

C1 — IMPACT BEFORE YOU FLIP THE SWITCH

AI tools often go live because someone saw a demo. Impact  before you launch. A one-page AI Impact Assessment asks five plain questions: benefits, harms, data, reviewer, and disclosure. It gives a Go / Hold / Conditions line.

​

​

​

DISCUSSION PROMPTS

1. Name one AI tool at work. What risk belongs on the table before it expands?

2. Who should sign off that impact was reviewed?

3. If it went wrong tomorrow, what disclosure would you wish you had written?

​

Checklist: C1-impact-assessment-checklist.pdf

​

C2 — WHO OWNS THE CALL

​

When AI writes, scores, screens, or recommends, a named human still owns the outcome. Use an ethical check before and after deployment, a clear escalation path, and never accept “the system said so” as a full answer.

​

DISCUSSION PROMPTS

1. For one AI-assisted task, who is responsible if it is wrong?

2. Where has “the system said so ended a conversation?

3. What would a 30-minute pre- and post-deploy check look like?

​

Walk away with an accountability map, named owner, escalation contact, vendor error path, and review dates.

Checklist: C2-accountability-map.pdf

​

C3 — ETHICAL STANDARDS AT WORK

​

“Be ethical” is not a standard. Use three rails: bias awareness, transparency, and honest labeling of AI assistance. If you cannot explain it, disclose it, or spot-check uneven treatment, it is not ready to ship as done.

​

DISCUSSION PROMPTS

1. Where could biased training data show up in your tools?

2. Can you explain one AI output in plain language?

3. When should AI assistance be disclosed?

​​

​

Walk away with a bias / transparency / labeling checklist, escalation language, and no-fake-human-claims rules.

Checklist: C3-ethical-standards-cert-lite.pdf

​

C4 — LEADERSHIP ONE-PAGER

​

Leaders need one page: policy → practice → measure. Use C1–C3 outputs as the executive view: what you require, what teams do weekly, and two metrics proving responsible AI practice.

​

DISCUSSION PROMPTS

1. What AI risk does leadership still underestimate?

2. What metric would prove responsible practice in 90 days?

3. Who owns the habit after this course?

​

Walk away with a one-page leadership roadmap, two metrics, a named owner, and a review date.

Checklist: C4-leadership-roadmap.pdf