I’ve been teaching at a nice resort. It’s hot as hell but hey, the aircon blows crazy cold and the kids are super. Debate for the next two weeks, love that topic.
Yet, I’m a little melancholic. It’s a good-bye for me. It is my last two weeks teaching (forever). It’s been a 35+ year journey and I’m grateful for how teaching has allowed me to live well, live energized, live engaged with thought and knowledge. But the moment has arrived. We’ll all (I hope) get there, to this point in life. I’m (re)tiring - putting on a new pair of tires.
I will keep writing here for the near term. I don’t know what I’ll do about the ELT Buzz Lesson Library. It’s not used much given the new obsession with “AI” so this too might have an “AI” sunset. Let’s see. I’m proud of it, a collosul lesson library +, but everything has its lifespan.
Soon, I plan on jumping on my bike to live my final years on the road. Getting out there and making love to the world. I’m riding off into the sunset.
I’ve had a good ride and feel honored to have taught and helped so many teachers enter and thrive in the hallways and classrooms of schools all over the world. I’ve fought what I think is “the good fight” - never selling-out and as much as I could, standing up for what I think is true and right in education.
I’ve seen a lot and participated in many educational domains. Technology being a primary one. I built ed-tech products and companies and trained lots of teachers. Larry Ferlazzo (another recently retired educator I respect immensely) even once offered a glowing endorsement.
So, with that in mind, my experience in ed-tech, I’d like to offer you dear reader, a brief summary of why I think education should reject “AI” except for true collating, sorting, statistical analysis, data work, word processing and typical “non-learning” tool type tasks.
I don’t think teaching/learning should welcome the embrace of an all-encompassing “AI” machine which makes everything seem so clean and direct when indeed, learning and intelligence itself is messy and oh so awry. Just say no.
In this Substack newsletter (now almost 5 years old), I’ve not written a lot about “AI”. However, find many articles, indepth on the topic, on my personal substack. Here, I’ve focused on practical teaching topics - what can help teachers directly in their classrooms. However, I’m going to make an exception today. This too, a last blast against the “AI” everything behemoth. So let’s begin …
Thoughts About AI & Why I’m An “AI” Refusnik.
It’s not “AI”. I have always been a stickler for accuracy in language terminology. I “…..” the acronym “AI” because it is inaccurate and actually a lie. Like calling a pet rock a dog. What we call and refer to as “AI” is really just super-charged word processing and “supervised” machine learning. Sure, it uses deep learning, sure it is generously coded in linear algebra. See this guy for an easy explanation. It has been around for awhile - we used to call it “big data”. But since “AI” got so much language data (theft of IP), it was able to add a front-end (a face) that allowed the average Joe and Jane to play with it. But it’s not “AI”, not intelligent - it’s still just a series of 1s and zeroes.
There are no shortcuts to learning. Students still have to do the work. An LLM generated output doesn’t teach, it doesn’t enable anything but the least bit of incidental learning. “AI” indeed leads to cognitive offloading, even surrender and a slide into dependency. Think for yourself, question technology. Write your own words. I have these words on my classroom walls.
Plagarism. Cheating. “AI” facilitates student cheating. And not all the focus on prompt engineering (moving the deck chairs on the Titanic) will change that fact. It is agency destroying. It is focused on producing an end-product, not co-creation (blablabla on that side-bar). Chomsky was correct, “AI” is all about skipping the hard stuff and the illusion of learning (see my next point). It’s a super plagarising machine. In fact, one may even go a step further — “AI” is itself cheating - stealing, remixing the knowledge of so many who came before. No citations, no acknowledgements — just theft. Faker. And not even doing it well. To end, let me ask a rhetorical question: Why do you think LLM generated text was not watermarked to be detectable? (and this could easily have been done, don’t believe those who say otherwise).
The Erosion of Authority. Everyone with “AI” is now an expert. We are turning away from those humans of us, who have years of experience in disciplines and just asking “AI” - LLMs. It’s become a rocket scientist in 15 minutes, time. And people believe that!!!! I bang my head daily when I hear people saying “Gemini said this …” Claude told me that …” Beyond the anthropromorphising - it is making an autocomplete function into an “expert”. And this is wrong, turning our backs on the human face of knowledge. Not to mention the bias, the errors, the hallucinations “AI” is prone to spit out. To use an LLM with any efficacy - you need to already be an expert. This is so true. And yet we are letting our young, still learning the basics, students use it? It’s like giving a blind man a kalashnikov.
Agency. Students using LLMs don’t search, don’t encounter “friciton” and they learn that knowledge isn’t something created, nurtured, cultured, of questions and “logos” even tacit but something that is impersonal, built, static, functionary. The UX of “AI” is set up to “humbug” and fool the user/learner with confident B.S. It is also designed to promote dependency, so you’ll return over and over, eschewing other sources of knowledge, knowing. All this notwithstanding that OpenAI has a marketing budget many times more than Coca-Cola! At bottom, education is about creating a hungry learner. You know the old saying - lighting a fire. Forming a student who is curious, hungry. You won’t do that by always throwing low-quality food into their dish.
A Litany of Human and Environmental Harms. Read this post for a rundown of these. There is little doubt about the harms of huge data centers to the environment we all live in. I’ve been to a large data center in Iowa. The noise itself is scary, freakish. “AI” is built on a very inefficient technology. And despite the “hype” and future wishing, that “AI” will become more efficient at scale - its inferencing takes enormous power to run, enormous resources. It will be a financial train wreck, mark my word. However the lies, the empty promises, the ethos of materialism and progress being technology won’t bear fruit. Rather they will lead us away from each other, destroy nature and allow control of the many, by the few. And that’s the problem in education. “AI” isn’t made with education or learning in mind. Learning wasn’t given a thought. It’s all about our data, digitizing everything so it can be turned into money. Student privacy - good luck! No way. Our souls are for sale when we use AI. And let’s not mention the dysfunctional way running to use “AI” for any answer — removes us from each other, discussing with each other, building knowledge with each other. That which is born is not that which is made.
Wisdom I believe is always circular. It’s not a straight, upward line. Samuel Butler in his Darwin Among The Machines advised us over 150 years ago to “not even go there” when it comes to defaulting to machine intelligence. I advocate that and ask you to do the same. Here is my reading of his timeless pamphlet.
Wish me good luck in my next venture, whatever that may be. I wish you all as I write this, this summer day, the same - with this one wild and precious life you have.
