Newsletter #5: The Blog hits 18K Views, Code Interpreter, AI Lie Detectors On Life Support, Creative AI, Learning Bots Grow, MasterDebater.ai, Higher Ed, More
Professor Mollick's crush on Code Interpreter; Darren Coxon's push; John Kelly's Future; Melissa McBride's Success; Laura Dumin's ideas; Shaping the Future With Dr. Sabba Quidwai
Stefan Bauschard
Educating4ai.com (Students grades 6-12, Cohort 1 starts this week); AI BootCamp (Adults, Enrolling Cohort II now). Contact stefanbauschard@globalacademic.org for institution-specific course adaptations and content licensing. Follow me on LinkedIn and check-out my co-edited 1,000 page volume on ChatGPT for educators.
Thank you!
“Education Disrupted” picked up 100 new subscribers this week and passed 18,000 views.
*AI Text Detectors
Damn, if this were a boxing match, I’d hope they’d just call it, because these detectors are really taking a beating. Check out Professor Mollick’s The Homework Apocalypse. Technology Review took their shot this week with “AI Text Detectors Are Really Easy to Fool.” This properly sums it up: Basically, it will catch the kid who copies and pastes the entire paper from a bot, but not one who makes any effort. Keep in mind that I was able to defeat a plagiarism detector with two words.
We are now at mockery stage –
And then there is good old-fashion plagiarism. Remember when kids used to copy and paste stuff from the internet and get caught by TurnitIn? Well, now they can run it through the Rewrite Plug-In, which will rewrite any page on the internet:
If they do this, they’ll have to use the normal tactics to defeat the plagiarism detectors after they are done, but they can do that quickly.
If I owned one of these companies, I’d be crying, Uncle! Or, I guess, I’d lie and swindle people, as I point out in my EdUp! Podcast with Jason Gulya. Yep, we have companies selling products based on lies whose entire business model is that it’s bad for kids to lie.
Sorry, the essay is mostly dead.
*Learning/Teaching Bots
Stuart Russell, one of the leaders in AI, says every kid in the world will have a capable teaching bots, probably in 2-3 years and no later than 2030. He points to studies that show 1:1 teaching creates learning gains that are 2-3 times as strong as classroom learning.
The Pi bot is really doing well.
And debaters can use MasterDebater.ai to practice against leading figures on their topics. It’s a great way to build an understanding of arguments.
As Darren Coxon notes, bots are also the only real hope for differentiation. Brian Cohen makes a similar point in, Tech is Once Again Disrupting Education.
Bots will be in elementary schools soon.
I also heard a few cool new ideas related to that: we’ll no longer need subject teachers; students won’t necessarily learn one subject at a time.
How are schools going to manage learning bots? Integrate them? Compete with them? Ignore them?
The Newark Public Schools experimented with Khamingo, the Khan Academcy AI bot, in May (Teachers Put A.I. Tutoring Bots to the Test) and the Australian government is testing EdChat in schools.
More: New A.I. Chatbot Tutors Could Upend Student Learning.
And we’re barely talking about what Brain Computer Implants mean for education….
*Code Interpreter
If you follow Ethan Mollick, you know he seems to have a crush on the Code Interpreter (CI) plug-in since he can’t ever stop talking about it. He had early access, but now everyone has it. As soon as that happened, he sent a Substack about it. CI is really useful. If you teach science, math, and economics, you and your students can have a lot of fun with this (and there are other creative uses for all). ‘Just don’t steal Code Interpreter from Professor Mollick; I think he’ll be crushed. I can’t blame him, CI is really all that and more.
Want to learn more about Code Interpreter? Sign-up for our AI Bootcamp!
*Creativity
One of the common criticisms of large language models is that they rely on “next word” prediction and that they are just repeating what has been said before rather than being creative.
Well, that’s basically BS. According to a new study.
This explains the comparative pace of progress in this area –
*Assessment
I’ve been saying this since at least May, but many people are starting to realize that what really matters right now is developing new assessments, and probably performative ones, because the papers and essays are meaningless. Dr. Phillipa Hardman has good stuff, and so does Sean McMinn. Also check out: Laura Dumin’s The case for changing How we approach grading and Embracing AI in Assessment. Laura also has a fun way to improve writing instruction using the detectors.
I’ve been working on suggestions for integrating AI into project based learning. Most of my background is in competitive debate; using debate in the classroom is coming soon.
One key question: What principles of instructional design assume a world where each student works with a Copilot/has their own 24/7 tutor?
*Push Push: More Educational Institutions Recognize the Significance of AI
Outside the international schools, it’s slow going, but many public schools are starting to pick up interest and are working hard to prepare for the AI World. I hope Japan’s integration of AI in schools gives more schools in the US a little push. It appears the ban on ChatGPT in Australian public schools will likely be overturned.
VR
No, it’s not dead, least of all in education. The first VR Charter School just completed its first year, and the VR High School, Sophia High School, has students who finished the year and crushed it academically. VR is coming to contemporary education. Check out their cool student presentations.
Tech Toys
Check out the combination of MidJourey and Runway.ml
*Disrupting Education
“This is the most disruptive thing ever..Don’t set homework essays.” Emad Mostaque, CEO of Stability of AI.
“ChatGPT4 can already reason better than many people” George Hotz,
“AI will change the course of history in the next 10 years.” Yuval Noah Harari
“Advancements in technology will bring personal computers on par with the processing power of the human brain. Coupled with virtual reality, brain mapping, and neural implants, these developments will enable the uploading of experiences and software directly into the brain.” John Kelly
Do you still think AI integration isn’t an academic emergency?
*Hands-On for Teachers
ScottyBreaksitDown (Australia). There are amazing hands-on AI tools here for teachers.
*Public Education
Phillipa Hardman’s The AI-Education Divide.Public schools are falling far behind private schools in AI education and training. This is a trend I’ve been observing since February, and given the importance of AI skills to the modern world of work, many students are suffering from a massive academic deficit (perhaps the most important one) that is not being measured. We do have a low cost program for students.
*Higher Ed
Jeppe Stricker’s Welcome to AI University
UK. universities draw up guiding principles on generative AI
US universities of the future: The University of Kentucky looks to welcome AI into Higher Education. Miami Dade College makes aggressive AI moves. There are plenty of ideas to replace schools and universities that don’t want to step-up.
*Tech Smaller Picture
How To Research Current Material on the Internet with ChatGPT
*Tech Big Picture
Peter Diamandis with Emad Mostaque (founder of Stable Diffusion): We’ve lost control. The big picture on AI and how we need to change our homework.
Lex Fridman with George Hotz: We might amuse ourselves to death. Current LLMs can reason better than many humans.
Will AI Save Education? Maybe Yes, Maybe No
Education is struggling (to say it kindly)
In the US, many teachers are quitting, and fewer people are choosing to become teachers. I hear reports about this from abroad as well. And test scores that measure national educational progress are declining. Many teachers report that students are disengaged. More and more students are medicated.
There is a compelling case to be made that the current system is broken and not preparing our students for the future (Manibot)
As a result, Darren Coxon doesn’t think the center will hold and that decentralized change is inevitably coming.
The answer? I think AI and its associated applications have great potential to support teacher retention and student engagement, but we have to be careful not to view these technologies as panaceas, as they have the potential to undermine teacher retention and leave our students vulnerable to algorithmic control.
What does AI mean for teacher retention?
Many think AI is the solution to the teacher retention crisis because it can save teachers time drafting quizzes, lesson plans, and such. There is truth to this, but it’s also going to add time to the teachers workday as they now have to learn the basics of AI, manage student “cheating,” and significantly adjust their teaching and assessment in the AI world, with little (it seems) support from administration. The teacher retention crisis is going to get worse unless administrators handle this transition to an AI-world really well.
This is my favorite thing about our AI Bootcamp: We not only teach administrators and teachers about the technology, but we also provide support for using Design Thinking to integrate the technology in a supportive and empathetic way. The tech is here and it’s not that hard to explain. Integrating it well is the challenge.
Will AI and PBL engage our students?
On the question of engagement, I’ve spent most of my career teaching students how to debate competitively. It’s highly engaging and involves little traditional learning. You know who 90% of the most committed debaters are? They are the same kids who sit in rows and crush it at rote learning in math and language classes.
AI-driven apps and a shift to project-based learning have the potential to engage more students, and maybe if we let the social media companies run the AI learning tools, many more students will be engaged, but these technologies aren’t easy solutions to the problems of student engagement and motivation. The also risk reproducing other forms of algorithmic social control, so we have to be very careful about how we implement them. Intentionality about what and how we are doing things is essential.
What are we trying to teach?
Content? Soft/Durable skills? A 50/50 mix? Can technology teach soft skills better than a human? Maybe AI tutors will be able to teach content better, but how important is content in an AI-World?
The future of education is uncertain, but there is no time like the present to start thinking about these questions. Others will ask (and answer) them if you don’t.
“The future of AI isn't something to fear, it's something to shape.”