What it is, how it works, and how marketers use ChatGPT and other tools.
robertbestwick.com/ai-digital-marketing/class-1/
Today
Agenda
By the end of class you can
Explain generative AI in simple words.
Describe how ChatGPT writes an answer.
Name 3 ways marketers use it, and 1 risk.
01
What is generative AI?
A short definition, a short history, and the key words.
Section 01
Three key words
Start with simple definitions.
1
AI
Computers doing tasks that usually need human thinking.
Example: a spam filter in your email.
2
Generative AI
AI that creates new content: text, images, audio, video or code.
Example: ChatGPT writes an ad.
3
Prompt
The instruction you type or say to the AI.
Example: "Write 3 Instagram captions for a café."
Sources: IBM, "What is generative AI?" (ibm.com/think); Google Cloud, "Introduction to Generative AI".
Section 01
A short history of AI
Click each year.
1950
Can machines think?
Alan Turing proposes a test: can a machine talk like a human?
Sources: Turing (1950); Dartmouth AI Project (1956); Vaswani et al., "Attention Is All You Need" (2017); OpenAI (2022, 2026); Reuters (Feb 2023); NobelPrize.org (2024).
Section 01
Why it matters now
Three numbers. Click each one.
Sources: Reuters, "ChatGPT sets record for fastest-growing user base" (Feb 2, 2023); OpenAI announcement (Feb 27, 2026); Statistics Canada, "AI adoption and productivity in Canadian firms" (2026).
Check
Four words to know
Tap each card.
Sources: IBM, "What are large language models?"; OpenAI Help Center, "What are tokens?"; Google Cloud, "What are AI hallucinations?"
02
How it works
No math. Just the big idea behind ChatGPT.
Section 02
From data to answer
Three steps happen behind every reply.
1
Train
The model studies billions of examples and learns patterns.
2
Prompt
You ask for something. Your words give it context.
3
Predict
It builds the answer one token at a time.
AI chat · illustration
Write a 1-line Instagram caption for a new Vancouver coffee shop.
Fresh beans, rainy-day warmth, and your new favourite corner in Vancouver.
Make it shorter and add a call to action.
Illustration based on a typical ChatGPT conversation. Source: Google Cloud Tech, "Introduction to Generative AI" (2023).
Section 02
It predicts the next word
An LLM picks a likely next token, again and again.
Our new coffee shop opens ___
Illustrative numbers only, to show the idea. Source: IBM Technology, "What are Generative AI models?" (YouTube).
Section 02 · Video
Watch: what are Gen AI models?
Listen for two words: “foundation model” and “risk”.
Video: IBM Technology, "What are Generative AI models?" (official IBM Technology YouTube channel).
Section 02
It predicts. It does not know.
Three limits every marketer must remember.
Hallucinations
It can invent facts, quotes and sources.
Bias
It learns from human data, so it can repeat human bias.
Privacy
Never paste private customer data into a free AI tool.
Rule: AI writes the first draft. You own the final version.
Sources: Google Cloud, "What are AI hallucinations?"; Office of the Privacy Commissioner of Canada, "Principles for responsible generative AI" (2023).
Check
Quick check
Tap an answer.
03
Traditional AI vs. generative AI
One predicts. One creates.
Section 03
Predict or create?
Click a row to compare.
Traditional AI
Generative AI
Main job
Predicts, sorts or recommends
Creates new content
Output
A label, a score or a choice
Text, images, audio, video, code
In marketing
Spam filters, ad bidding, product recommendations
Ad copy, product photos, chatbot replies
Together: traditional AI picks who sees the ad. Gen AI helps write what it says.
Sources: IBM, "AI vs. machine learning vs. deep learning vs. generative AI"; Google Ads Help, "About Smart Bidding".
Section 03 · Demo
Before Gen AI: rule bots
Meet CC-Chatty. Type a question, or tap one below.
IF message has
THEN reply
1
schedule, class
Your class schedule is on the LMS.
2
assignment, submit
Upload assignments to the LMS as a PDF.
3
attendance, absent
You need 80% attendance in each course.
4
hello, hi
Hi! Ask me about classes or assignments.
–
ELSE
Sorry, I don't understand.
CC-Chatty · class demo
Class demo, based on the CC-Chatty student assignment. Rule 3 uses the attendance rule in the DMM-260 Learning Plan 2026.
Check
Sort it
Drag each card, or tap a card and then tap a box.
Gmail filters spam emails
Writes 5 Instagram captions
Recommends your next Netflix show
Creates a new background for a product photo
Predicts which customers may cancel
Drafts a welcome email
Traditional AI predicts
Generative AI creates
04
Real-world applications
What it makes, where it adds value, and what can go wrong.
Section 04
What Gen AI can make
Weeks 4 and 5 go deeper into each type.
Text and code
Posts, emails, blog drafts, simple websites.
ChatGPT, Claude, Gemini, Copilot
Images
Product scenes, ad visuals, mood boards.
Adobe Firefly, DALL·E, Midjourney
Audio and video
Voice-overs, music, short video clips.
ElevenLabs, Google Veo, OpenAI Sora
Sources: product pages of each tool (openai.com, anthropic.com, gemini.google.com, adobe.com/firefly, elevenlabs.io, deepmind.google/veo).
Section 04
Where the value is
McKinsey studied 63 business uses of Gen AI.
$2.6–4.4T
possible value per year for the global economy
About 75% of that value is in 4 areas:
Customer operations✓ Marketing and salesSoftware engineeringResearch and development
Marketing is one of the biggest winners. That is your field.
Source: McKinsey & Company, "The economic potential of generative AI: The next productivity frontier" (June 14, 2023). T = trillion US dollars.
Section 04 · BC case
When the chatbot is wrong
Moffatt v. Air Canada, BC Civil Resolution Tribunal, 2024.
1
The mistake
Air Canada's website chatbot gave a customer wrong information about a bereavement (family death) discount.
2
The defence
The airline said the chatbot was "responsible for its own actions".
3
The decision
The tribunal said no: the company owns everything on its website. Air Canada had to pay.
Source: Moffatt v. Air Canada, 2024 BCCRT 149 (February 2024); McCarthy Tétrault, "Moffatt v. Air Canada: A Misrepresentation by an AI Chatbot" (2024).
05
Generative AI in marketing
Two famous campaigns, three daily uses, three rules.
Live action, effects and AI together. Praised by viewers.
2
2024: AI holiday ad
Made mostly with AI. Many viewers called it cold.
3
The lesson
People judge the feeling, not the tool.
Sources: Coca-Cola (official YouTube channel); Creative Bloq, "Coca-Cola's new ad is a branding masterpiece" (Mar 2023); NBC News, "Coca-Cola causes controversy with AI-made ad" (Nov 2024).
Section 05
How marketers use it every day
You will practise all three in this course.
1
Create
First drafts of posts, emails, ads and images.
Weeks 2, 4, 5
2
Personalize
One message, many versions for different audiences.
Week 10: email campaign
3
Research and plan
Customer personas, competitor summaries, SEO ideas.
Weeks 3 and 8
SEO = search engine optimization: helping a page appear higher on Google. Source: DMM-260 Learning Plan 2026.
Section 05
Three rules for this course
And for your future job.
Human in the loop
Always read, check and edit before you publish.
Protect data
No private client or customer data in free tools.
Be transparent
Say when AI made an image or a big part of the work.
Sources: Office of the Privacy Commissioner of Canada, "Principles for responsible, trustworthy and privacy-protective generative AI technologies" (Dec 2023); Canadian College Cheating & Plagiarism Policy.
Check
Quick check
Marketing edition. Tap an answer.
In-class assignment
Design your own IF/THEN bot
Build a rule-based chatbot, test it, and compare it with Gen AI.
1
Choose a topic and a name
Student life or a local business you know.
2
Write 5 rules + 1 ELSE
IF the user types a keyword, THEN your bot replies.
3
Test, reflect, upload
Try 3+ messages, fix your rules, download the PDF.