AI In The Loop.
Teacher In Charge.
Bellringer
1. Answer the question, turn your paper upside down, pass to one person near you…then pass it again.
2. Everyone should have a paper, nobody should have their own.
3. When I say go:
4. When the music stops, head back to your seat.
The AI Enthusiasm Spectrum
AI will save us
AI will destroy us
AI’s okay.
risk ignoring the flaws
Goal today: Join me in radical centrism.
risk ignoring the opportunities
Goal for today (keynote + workshops)
Leave you feeling in full control of AI in your classroom.
Goal for the keynote
Basic understanding of AI & centrist view on its potential.
“Teacher in the loop”
AI in the loop, teacher in charge.
1. AI Basics
Today
2. The risk
3. The opportunity
4. The Path Forward
Part 1: Understanding AI
First, a toothbrush game
AI is the future of work. It can already….
Enable humans to control computers with their brains
Conduct open ended tasks on the web.
Serve as a rudimentary family therapist.
Beat Grandmasters of Chess and Go.
Use K-Cups to make coffee and reason through physical tasks.
Win art competitions.
Generate videos with basic text prompting.
Replicate human thought, speech, and behavior.
“This is the worst AI we will ever use.”
AI is the future of work (and maybe learning)
AI is the future present
AI is the future present. Students are using it to….
Augment search results in Google
Embedded in the Iphone
In Snapchat & Twitter
Writing emails
Writing docs
Filling out forms in Chrome
Writing text messages
In Google Classroom
93% of students have used AI “once or twice” to help with schoolwork
AI is the future present.
We’re doing a disservice to students if we don’t teach them how to use AI tools effectively.
…and sometimes that means teaching them when not to use AI at all.
What are LLMs…
LLMs predict the next “token”
Foundation Model
Foundation Model
token
20%
10%
98%
Word associations based on training data (Reddit, Wiki, Twitter, etc)
The text generated when you “talk” with ChatGPT
LLMs
Foundation Model
20%
10%
98%
LLMs do not “think”, they imitate thought.
…if you can’t tell the difference, does it matter?
LLMs do not “think”, they imitate thought.
LLMs with more data are better at imitating thought.
Foundation model with less/worse data
Qwen 2.5 Coder: “Write me a sonnet in the style of Matthew McConaughey”
Upon the moors where echoes softly call,
Beneath the vast and indifferent sky so wide,
A lone soul wanders, lost in reverie,
Where shadows stretch and time seems to abide.
The heather sways in gentle, mournful dance,
A symphony of nature’s breath and sigh,
Each tuft a whisper, secrets it retains,
Of ancient storms and silent winters by.
20%
10%
98%
Model with more/better data
Claude 3.5 Sonnet: “Write me a sonnet in the style of Matthew McConaughey”
The world keeps turnin', that's what I say,
Like a slow dance beneath the Austin stars.
Some folks get lost along their own highway,
Too busy countin' miles, collectin' scars.
But here's the truth I found while drivin' far:
Joy ain't about the destination, see?
Just like that Lincoln, cruisin' in my car,
The journey's what made sense of life to me.
The universe, she whispers ancient tales,
Of time that flows like rivers through the night.
When wisdom comes, your spirit never fails,
You're just L-I-V-I-N', and that's alright.
So keep on rollin' through this cosmic scene,
'Cause life's a wave, you know just what I mean.
Models (i.e. LLMs) produce different outputs depending on the amount and quality of their training data.
LLMs in Review
Learn more
Part 2: The risks
The amount and quality of LLM training data can lead to negative side effects…
The AI Enthusiasm Spectrum
AI will save us
AI will destroy us
AI’s okay.
Join me!
Hallucination
missing data
Hallucination
“What happened on September 11th, 2001?”
Trained on structure AND content. Can answer this accurately.
“What happened on February 12th, 1249?”
Familiar with structure, but NOT content, making it more likely to produce false info.
structure
content
Hallucination consequences
Unintentional Bias
Prompt
“Help me generate a story about a powerful leader who helps a country grow.”
“man”
“powerful”
Intentional Bias
Intentional Bias
IP Theft?
IP Theft?
IP Theft?
VS
Carbon Footprint
The devaluing
of literacy.
Because of AI, humans will read less.
Because of AI, humans will write less.
We’re writing less
Losing our voices
We’re losing our voices.
Hallucination
Unintentional
bias!
Intentional
bias!
Less writing!
Less reading!
Homogenization of voice!
The AI Enthusiasm Spectrum
AI will save us
AI will destroy us
AI’s okay.
The AI Enthusiasm Spectrum
AI will save us
AI will destroy us
AI’s okay.
Part 3: The opportunity
“I want AI to do my laundry and dishes so that I can do art and writing, not for AI to do my art and writing so that I can do my laundry and dishes.”
AI can’t do your dishes (yet). But…
AI saves teachers 5.9 hours per week…6 weeks per year.
40-50% of teachers leave the profession within the first 5 years.
Efficiency.
Efficiency
“We find that the current iteration of ChatGPT scoring is not statistically significantly different from human scoring”
Efficiency
“Differences between ChatGPT and humans were modest when considering the overall quality of feedback and time-savings…”
No more Sunday scaries!!
Personalization
Personalization
Personalization
Personalization of literacy.
AI is opening up new roads to literacy.
New roads to literacy
“Living books”
AI is devaluing literacy
AI is enabling literacy.
Part 4: Path Forward
The AI Enthusiasm Spectrum
AI will save us
AI will destroy us
AI’s okay.
“Don’t call it a sign of the times…
…When it’s always been this way.”
470 BCE!
The literacy enthusiasm spectrum
Literacy will destroy us
Literacy will save us
Literacy itself might be our best “AI Analogy.”
470 BCE!
Written language is dangerous!!
Writing … “drifts all over the place, getting into the hands not only of those who understand it, but equally of those who have no business with it! It doesn’t know how to address the right people, and not address the wrong!”
- Socrates
Hol’ Up.
Plato
Writing allows us to preserve knowledge accurately over space and time! It enables modern history, religion, science, literature, and law!
Plato
We can scale our expertise! Written texts can be duplicated and distributed!
Plato
Individuality! One can wrestle with ideas on their own for the first time!
Socrates was…
kind of right.
Modern Socrates
Plato was…
also kind of right.
Modern Plato
The literacy enthusiasm spectrum
Literacy will destroy us
Literacy will save us
Literacy’s pretty cool
Technology amplifies human intent.
No one has more power to shape human intent than teachers.
AI makes teachers more important, not less.
Society
Teachers
Our work starts
now.
Thank you.
adam@myshortanswer.com - 402 657 8205
Slides!
bit.ly/esc3literacyrules