# Quick-Reference Card: AI Foundations for Every Employee

<p class="meta">FAST-15 · 7formations · Instructor: Wadood Chaudhary · Keep at your workstation · Based in part on Svekis, AI Literacy for All Employees (2026), chapters 1 to 5</p>

## 1. Is it AI?

| Ask | If yes |
| --- | --- |
| Did a person write the rule? | Rule-based automation (not AI) |
| Did it learn from data to estimate an outcome or category? | Predictive AI |
| Does it create new text, images, or audio? | Generative AI |

## 2. How it learns

**Supervised** (labeled examples) · **Unsupervised** (finds groups) · **Reinforcement** (rewards after actions) · Chat assistants: predict next words on huge text, then tuned by people.

## 3. Why it can be wrong

Limited data · Outdated information · Misread context · Bias in data. **Probable is not certain.**

## 4. Spot hidden AI

Features that **suggest, predict, summarize, rank, generate, transcribe.** Ask: What does it do? What data does it see? What do I still decide?

## 5. Myths to drop

AI thinks like us · is always right · can replace judgment · gets context · is objective · learns from (or forgets) every chat. **Check your tool's memory and data settings.**

## 6. The assistance ladder

| 1 Inform | 2 Suggest | 3 Draft | 4 Act with approval | 5 Act alone |
| --- | --- | --- | --- | --- |
| You read | You choose | You edit and own | You approve first | Limits plus review |

Higher rung = more limits and review. Decisions about people's jobs, pay, or benefits need a named person.

## 7. Data rule

Never paste name + SSN, license or state ID, or account or card number (G.L. c. 93H, 201 CMR 17.00). Use approved tools only.

*Approved tool(s): ____________  ·  Ask or report to: ____________*