# Session 2 Handout: AI in Everyday Work and the Myths Around It

<p class="meta">FAST-15 AI Foundations for Every Employee · 7formations · Instructor: Wadood Chaudhary · All examples fictional · Suggested reading: Svekis, AI Literacy for All Employees (2026), chapters 3 and 4</p>

## Why this session matters

Many employees say "we don't use AI" while their email, phones, and planning tools use it every day. Hidden AI still makes mistakes and still sees data. And beliefs about AI, both over-trusting and over-fearing, drive real behavior. This session helps you spot AI where it already is and replace six common myths with accurate statements.

## Key concepts

### 1. Five categories of workplace AI

| Category | Examples |
| --- | --- |
| Content and communication | Suggested replies, drafting, translation |
| Information processing | Summaries, transcription, search |
| Decision support | Fraud alerts, demand forecasts, route suggestions |
| Workflow automation | Routing tickets, extracting invoice fields |
| Personal productivity | Scheduling help, note-taking, task lists |

### 2. Automation versus augmentation

**Automation:** the AI does a step on its own (a fraud system declines a card charge). **Augmentation:** the AI helps a person do the step (a dispatcher sees a suggested reroute and decides). Most workplace AI today is augmentation, and the person stays responsible.

### 3. How to spot hidden AI

Look for features that **suggest, predict, summarize, rank, generate, or transcribe**. For each one, ask three questions: What does the AI do? What data does it see? What does a person still decide?

### 4. Six myths and the reality

| Myth | Reality |
| --- | --- |
| AI thinks like a person | It finds statistical patterns; it has no understanding or awareness |
| AI is always accurate | It can be wrong, outdated, or incomplete, and sounds just as sure when it is |
| AI can replace human judgment | It lacks ethical reasoning and full knowledge of your situation |
| AI understands context perfectly | It misses context it was not given: local rules, recent changes, unwritten norms |
| AI is objective | Its data and design can carry bias |
| AI learns automatically from every chat (or "nothing I type is kept") | The model is not retrained by your chat, but some tools remember details across chats, and some plans let the vendor use chats to train future models. Check your settings and policy |

The book (Svekis, chapter 4) explains the first five myths well. For myth 6, the course adds 2026 detail: OpenAI's help center, for example, describes memory you can turn off and says business plans are not used for training by default, while consumer settings may be. Other vendors have similar pages.

## Worked example

Merrimack Freight & Logistics (fictional) turns on an AI meeting recap. The recap says "driver pay increase approved," but the call only discussed it. The transcript also captured a driver's medical update. Two lessons: built-in AI output gets the same check as chatbot output, and transcripts can hold sensitive information, so follow your employer's rules on who can see them.

## Try this before session 3

- Open the settings page of one AI feature you use and find its memory or data controls.
- Read Svekis chapters 3 and 4.