Dancing with the Devil We Know: OpenAI and the Future of Education
Analyzing OpenAI's Student Writing Guide and Latest AI Tools
🎁 Special Holiday Offer! 🎁
Subscribe to my premium content before December 31st and lock in a 20% discount forever. That's right - you'll keep this special rate for as long as you remain a subscriber!
Don't miss this limited-time opportunity to join our growing community of forward-thinking educators at a permanent discount.
Personal Evolution with AI
In the summer of 2024, I found myself on something of a tear. OpenAI had just released their ChatGPT 4o model, and I spent weeks hammering away at posts about their multi-modal products – tools they had essentially unleashed on the world (read: our students) without so much as a safety net in sight. The more critically I wrote about their impact on education, the more my engagement soared. It seems I wasn't alone in my unease about technology that required nothing more than a phone number and a willingness to hand over your work (temporarily, they claim, unless you forget to uncheck that sneaky 'share for improvement' box).
Now, with last week's launch of the full GPT o1 Pro version, I find myself reflecting on how my perspective has evolved. Back in the summer, watching the demo release felt like seeing the walls of Jericho come tumbling down. In my most paranoid moments, I even caught myself thinking of OpenAI's new multi-modal tools as some kind of AI-driven TikTok 2.0. But that metaphor wasn't quite right. What I've come to understand is that AI's influence isn't manifesting through dramatic upheaval, but through what I think of as ambient presence.
Despite all the breathless hype about AI developing agency, these tools remain surprisingly bound by rules and limitations. Yet their influence seeps into our work in ways we might not always notice. As I use AI daily – more Claude these days than GPT – its generative properties form a kind of medium I operate within. Even as I push back against sub-standard responses and carefully reword prompts to keep interactions on track, these models exert a kind of gravity. There are biases and assumptions that I pragmatically choose not to fully engage with in pursuit of more immediate payoffs.
OpenAI's Recent Developments
Since that summer, I've backed away from my strong anti-OpenAI posturing. Notably, my article engagement has dropped – I guess people don't get quite as excited about practical solutions and even-handed engagements with the larger AI landscape. But I'm at peace with this repositioning. I find I have more energy now to focus on what really matters: curriculum building, instructional development, and honest assessment of what's happening in the AI corporate universe. This is, after all, the core of my work as an educator and AI consultant building relationships with a growing network of schools in my area and beyond.
Today, I want to reflect on two recent OpenAI developments that highlight this evolution: their belated publication of advice for students on integrating AI into writing workflows, and last week's launch of the full GPTo1 Pro version. When OpenAI released their student writing guide, there were plenty of snarky comments about how this guidance arrives almost a year after they thoroughly disrupted the educational landscape. Fair enough – I took my own side swipes initially. But let's look at what they're actually advising, because the details matter more than the timing.
Analyzing OpenAI's Educational Guidance
Let's look at these recommendations through the lens of practical classroom reality rather than corporate idealism:
The Good: Process and Critical Thinking OpenAI gets it right when they focus on writing as a process. Their suggestions about using AI for structural feedback, reverse outlining, and Socratic dialogue are genuinely useful tools – areas where ChatGPT truly excels. When they encourage students to "pressure-test" their thesis statements with counterarguments, they're promoting exactly the kind of critical thinking we want to develop.
But here's the catch: all of these tools require a foundation. Students need to know how to outline independently before using AI to critique their structure. They need experience in real human Socratic dialogue before engaging with an AI version. They need to understand how to craft a thesis statement before asking for AI-generated counterarguments. The technology isn't a shortcut – it's an amplifier of existing skills.
The Bad: Research and Citations Where OpenAI seriously misses the mark is in their advice about citations and research. Their suggestion to "delegate citation grunt work to ChatGPT" is particularly problematic. In my experience, ChatGPT remains unreliable for citations – traditional tools like EasyBib are still far superior. Even more concerning is their recommendation to use AI for "getting a roadmap of relevant sources." When I test ChatGPT's source recommendations, I'm finding roughly 50% are still hallucinated. Yes, the model will usually admit this if directly questioned – but how many students will think to ask?
The Uncertain: Reading and Learning Some recommendations sit in a grey area. Take their suggestion to "use Advanced Voice Mode as a reading companion." For some students, particularly those who struggle with reading comprehension, this could be valuable support. But I worry about the majority of my students who already struggle to maintain focus for even 5-10 minutes of sustained reading. Adding AI commentary might create more distraction than assistance.
The Reality Check: Here's where OpenAI and I find common ground: their final point about not just "going through the motions" but using AI to develop critical thinking and clear writing skills. This is exactly right – but it vastly underestimates the human infrastructure needed to make it work. We need supportive resources, curricular vehicles, teacher training, and carefully designed student routines. These tools won't magically transform education by themselves.
GPT o1 Pro: OpenAI's Business Evolution
The recent launch of GPT o1 Pro signals another shift in OpenAI's evolution – one that confirms what I've suspected since tracking the deceleration of scaling laws: OpenAI is no longer primarily a breakthrough company. They've become a products company. The products are impressive, yes, but at current price points, not particularly enticing. It feels like we're watching a company trying to recoup massive investments by scaling up into business markets, finally hoping to extract real value from their enormous expenditures in research, design, training, and ongoing marketing spectacles.
Will it pay off? Perhaps. Unlimited access does carry real value. But GPTo1 Pro is definitely not GPT-5 – it's more like GPT-4 with some fine-tuning. OpenAI must tweak and squeeze, letting their models think for longer intervals just to see marginal improvements in math and coding capabilities. And anyone who has used o1 Pro for writing knows that GPT-4 and the Claude suite still offer superior performance.
Yet here's the paradox that keeps educators up at night: OpenAI continues to build state-of-the-art AI technology that forms the foundational architecture for countless AI tools. Even when schools work through protective AI layers, OpenAI's technology tends to be powering things under the hood. Meanwhile, students continue to use OpenAI products outside of course routines – sometimes in amazingly productive ways, sometimes simply to maximize efficiency. This is the complicated legacy of birthing a largely unregulated general application, and what makes their delayed educational guidance sting so sharply.
Image generated by Napkin.ai
The OpenAI Strategy
Perhaps this is exactly where OpenAI wants us – caught in their orbit, alternating between criticism and fascination. They seem uninterested in playing the role of the sedate, dependable second player in the market. Instead, their strategy appears to be: start fires, push boundaries, make audacious claims, occasionally erode trust, but keep dropping next-level tools first. The result? Even as you critique them, you find yourself tempted to resubscribe, just to see what's next.
I'll admit it – they've pulled me through this cycle more than once. Right now, I'm finding enough value in o1 Pro to maintain my subscription. But for how long? Claude has me thoroughly impressed at the moment. Yet even this preference reveals another trap of American tech consumerism: this company or that company, this subscription or that subscription, always seeking the next best thing.
Navigating the Educational Landscape
The reality for educators is more complex than choosing sides. We're tasked with preparing students for a world where these tools are ambient, omnipresent. The challenge isn't just teaching with or about AI – it's helping students develop the critical awareness to navigate this new landscape while maintaining their agency. OpenAI may have launched their educational guidance late, but they've inadvertently given us something more valuable: a real-world case study in how technology reshapes educational landscapes, for better or worse.
As I continue building curricula and training teachers, I find myself returning to a simple truth: our role isn't to be for or against these tools, but to help students understand their gravity. Like the internet before it, AI isn't just a tool – it's a medium that shapes how we think, write, and learn. The sooner we acknowledge this reality, the better equipped we'll be to help students maintain their intellectual independence while harnessing AI's potential.
For now, like many educators, I'll keep one foot in OpenAI's ecosystem while maintaining my critical distance. It's an uneasy balance, but perhaps that's exactly what teaching in the AI age requires: the ability to embrace utility while resisting complete absorption into any one company's vision of the future.
Nick Potkalitsky, Ph.D.
Transform Your School's AI Culture
Two Pathways to AI Excellence:
School Package: Enroll your team, get a free AI workshop Free strategy sessions & implementation support available.
Transform your school's AI readiness: nicolas@pragmaticaisolutions.net
Check out some of my favorite Substacks:
Mike Kentz’s AI EduPathways: Insights from one of our most insightful, creative, and eloquent AI educators in the business!!!
Terry Underwood’s Learning to Read, Reading to Learn: The most penetrating investigation of the intersections between compositional theory, literacy studies, and AI on the internet!!!
Suzi’s When Life Gives You AI: An cutting-edge exploration of the intersection among computer science, neuroscience, and philosophy
Alejandro Piad Morffis’s Mostly Harmless Ideas: Unmatched investigations into coding, machine learning, computational theory, and practical AI applications
Michael Woudenberg’s Polymathic Being: Polymathic wisdom brought to you every Sunday morning with your first cup of coffee
Rob Nelson’s AI Log: Incredibly deep and insightful essay about AI’s impact on higher ed, society, and culture.
Michael Spencer’s AI Supremacy: The most comprehensive and current analysis of AI news and trends, featuring numerous intriguing guest posts
Daniel Bashir’s The Gradient Podcast: The top interviews with leading AI experts, researchers, developers, and linguists.
Daniel Nest’s Why Try AI?: The most amazing updates on AI tools and techniques
Riccardo Vocca’s The Intelligent Friend: An intriguing examination of the diverse ways AI is transforming our lives and the world around us.
Jason Gulya’s The AI Edventure: An important exploration of cutting edge innovations in AI-responsive curriculum and pedagogy.
Yup, I also arched my eyebrows when I read the advice to "Use ChatGPT for citations." Knowing how often LLMs make up plausible-sounding sources and links, that's just bad advice, especially if the person reading it isn't aware of hallucinations.
They try to paint over it with "double-check the sources," but that would require exactly the kind of "grunt work" that they suggest replacing.
I agree that ChatGPT can help with citation formatting and structuring, but you still gotta do your legwork to make sure it's not all bullshit.
I appreciate your nuanced take, even if it's less effective at driving clicks.
Your conclusion is spot on. In my conversations with students, while getting a good grade is always present as a motivating factor, students *also* know they're going to have to use AI in their careers. They want to begin exploring AI usage within the "safe" confines of school, in which a trusted teacher can help them learn how to use it effectively. Unfortunately, AI is still largely forbidden (for good reasons!) in academic settings, which means student usage of AI is guided by what they see on TikTok.