Introduction to AI

AI in the
Professional
Workplace

A practical guide for working professionals — no technical background required

What is AI? How LLMs Work Key Tools Hallucination Responsible Use Productivity Ethics Interactive Q&A
Dane Adams  ·  daneadams.co.uk
01 / 14
Context

Why AI Literacy
Matters Now

AI is not a future trend. It is already embedded in tools professionals use every day — from email to search to document creation. Understanding it is rapidly becoming a baseline professional skill.

92%

of CHROs expect AI to be further integrated into their organisations in 2026 (SHRM 2026)

54%

of organisations have not yet adopted AI and have no current plans to — a rapidly closing window (SHRM 2026)

120+

court cases logged where AI hallucinated fake legal citations — submitted by professional lawyers (2025–26)

The professionals who thrive are not necessarily those who know how to build AI — they are those who know how to use it effectively and safely.

02 / 14
Foundation

What Is AI,
Actually?

Artificial Intelligence is the broad field of building systems that can perform tasks that would normally require human intelligence — recognising speech, making decisions, generating text or images.

Generative AI is the specific type most relevant to professionals today. It creates new content — text, images, code, audio — in response to a prompt.

The tools you will use daily (ChatGPT, Copilot, Claude) are all generative AI. They are powered by Large Language Models (LLMs).

The Simple AnalogyImagine reading every book, article, email and webpage ever written. After all that reading, you become extraordinarily good at predicting what word comes next in any sentence. That is what an LLM does. The critical implicationIt predicts. It does not know. This is why it can be wrong — even when it sounds very confident.
03 / 14
Tools

The Tools You
Need to Know

ChatGPT — OpenAI

The most widely used AI assistant. Excellent for drafting, summarising, reasoning, and coding. GPT-4o is multimodal — it handles text, images, and voice.

Microsoft Copilot

AI built directly into Microsoft 365. Drafts emails in Outlook, summarises meetings in Teams, analyses data in Excel, and generates presentations in PowerPoint.

Claude — Anthropic

Strongest for long-form writing, nuanced editing, and working through lengthy documents. Known for following complex instructions carefully.

Midjourney / DALL·E

Image generation tools. Create visuals, concept art, and design assets from text descriptions. Increasingly used in marketing, comms, and design workflows.

Key principle: Different tools have different strengths. The right tool depends on your task — not brand loyalty.

04 / 14
Quick Check

Question 1

A colleague tells you that ChatGPT "knows" all the answers because it was trained on the entire internet. What is the most accurate response?

AThey are correct — the model has memorised all that information and retrieves it accurately.
BThey are mostly right, but the model only knows things up to its training date.
CThe model predicts likely responses based on patterns — it does not retrieve or verify facts, which is why it can be confidently wrong.
DIt depends on which version of ChatGPT they are using.
05 / 14
Critical Concept

Hallucination —
What It Is & Why
It Matters

Hallucination is when an AI model produces confident, fluent, completely incorrect output. It does not know it is wrong. It has no concept of truth — only of likely text.

This is not a bug that will be fixed. It is a fundamental property of how LLMs work.

Real-world consequencesLawyers submitted court documents citing cases that did not exist — generated by ChatGPT. 120+ documented cases. The EU AI Act now imposes penalties up to €35M for AI-related harms.
1

Never copy AI output directly

Treat every AI response as a first draft that requires human review — especially facts, figures, names, and citations.

2

Ask for sources separately

If you need a citation, find it yourself. AI-generated citations are frequently fabricated and formatted convincingly.

3

Be more sceptical when it's most confident

Hallucinations are most dangerous when they sound authoritative. Certainty is not accuracy.

06 / 14
Quick Check

Question 2

You use ChatGPT to draft a market research report and it cites three industry studies with author names, dates, and page numbers. What should you do?

AInclude them — the model was trained on vast amounts of research literature so they are likely real.
BInclude them but add a disclaimer that they were AI-generated.
CVerify every citation independently before including any of them — AI-generated citations are frequently fabricated.
DRemove the citations and ask the AI to regenerate them without page numbers, as those are more likely to be accurate.
07 / 14
In Practice

Where AI Adds
Real Value at Work

Drafting & Writing

First drafts of emails, reports, and documents. Editing for tone and clarity. Adapting content for different audiences. Always edit the output.

Summarising

Condensing long documents, meeting transcripts, or research into key points. Excellent for processing information quickly — verify critical details.

Brainstorming

Generating options, angles, and ideas rapidly. AI excels at breadth — use it for ideation, then apply your own judgement to select and refine.

Data & Analysis

Interpreting data, writing Excel formulas, explaining outputs. Copilot in Excel can analyse spreadsheets and answer questions about your data directly.

Research Starting Points

Explaining concepts, mapping a topic area, generating questions to investigate. Never end your research with AI — use it to begin, not conclude.

Workflow Automation

Templating repetitive tasks, generating standard communications, structuring processes. Significant time savings on high-volume, low-complexity work.

08 / 14
Limitations

Where AI Falls
Short

Current information

Most models have a knowledge cutoff. They do not know about last week's news, updated legislation, or new product releases unless connected to live search.

Verified facts and citations

As covered — AI generates plausible-sounding content, not verified information. Never trust a statistic or citation without checking it.

Your organisation's internal context

AI does not know your company, your clients, your processes, or your confidential information — unless you tell it. And telling it carries data privacy risks.

Nuanced human judgement

Relationships, politics, ethics, context, timing — AI lacks the professional and emotional intelligence to navigate these reliably.

Consistency across sessions

AI has no memory between conversations by default. It will not remember decisions made last week, or context from a previous session.

Replacing professional accountability

You are responsible for what you submit — whether AI helped write it or not. The courts have been clear: AI does not reduce professional liability.

09 / 14
Quick Check

Question 3

A team member wants to paste a confidential client proposal into ChatGPT to get help improving it. What is the main concern?

AChatGPT might not understand the technical content of the proposal.
BConfidential information entered into a public AI tool may be used for model training or exposed in a data breach — creating a serious data privacy risk.
CThe suggestions AI makes may not match your organisation's brand voice.
DThere is no real concern — public AI tools are designed for professional use.
10 / 14
Ethics & Governance

Using AI
Responsibly

1

Verify before you publish

Every AI output that will be shared externally needs human review. You are accountable for what goes out under your name or your organisation's name.

2

Protect confidential information

Do not enter client data, personal data, or proprietary information into public AI tools. Check your organisation's policy first.

3

Be transparent when it matters

In many professional and academic contexts, disclosing AI use is expected or required. Know the rules that apply to your situation.

4

Watch for bias

AI models can reflect and amplify biases present in their training data — particularly in hiring, assessment, and communication contexts. Apply critical thinking.

5

Keep humans in the loop

For decisions with significant consequences — personnel, legal, financial — AI should inform and assist human judgement, not replace it.

6

Stay current

AI tools and regulations are evolving rapidly. The EU AI Act is now in force. What is acceptable practice today may change. Stay informed.

11 / 14
Quick Check

Question 4

Which of the following is the best description of how AI should be used in a professional context?

AAI should make the final decision in any situation where it has sufficient data — it removes human bias from the process.
BAI should augment human judgement — handling repetitive, time-consuming tasks while humans retain responsibility for decisions and outputs.
CAI is only useful for creative tasks like writing and image generation — it adds limited value to analytical professional work.
DAI use should be avoided until regulations are clearer, to reduce organisational risk.
12 / 14
Next Steps

Getting Started
This Week

1

Pick one task you do repeatedly and try AI on it

Draft an email. Summarise a document. Generate an agenda. Start with something low-stakes where the cost of a mistake is minimal.

2

Be specific in your prompts

Tell the tool who you are, what you need, who the audience is, and what format you want. Vague prompts produce vague outputs.

3

Treat the first output as a draft, not a result

Ask follow-up questions. Refine iteratively. The best outputs come from a conversation, not a single prompt.

4

Check your organisation's AI policy

Before using AI tools at work, understand what is and is not permitted — particularly regarding confidential data and client information.

5

Stay curious — the tools are evolving rapidly

The best way to keep pace is through regular use, not occasional reading. Build a habit of experimentation and critical evaluation.

13 / 14
Summary

What We
Covered Today

What AI Is

LLMs predict likely text — they do not retrieve or verify facts. Understanding this changes everything about how you use these tools.

Key Tools

ChatGPT, Copilot, Claude, Midjourney. Different strengths — choose based on the task, not habit.

Hallucination

AI can be confidently wrong. Verify facts, never trust citations without checking, and always review before publishing.

Where It Adds Value

Drafting, summarising, brainstorming, data analysis, workflow automation. Strong for volume — always needs human review.

Responsible Use

Protect confidential data. Be transparent. Watch for bias. Keep humans accountable for decisions. Follow your org's policy.

Getting Started

Pick one task, be specific, iterate, and verify. Build the habit through regular, critical, low-stakes experimentation.

AI will not replace professionals who use it well. It will replace professionals who ignore it entirely.

14 / 14
1 / 14

← → arrow keys to navigate  ·  F for fullscreen