DebateUS AI Education Newsletter #4!
Newsletter Round-Up
*Tech keeps advancing and getting better
*Deep Fake Trouble
*Conceptual and Practical Uses in Education
*Web Extenders
*Ethics & Unemployment
*The End of the World?
*Global Reaction
Things Keep Getting Better
If I've had one theme since I started working on this in January, it's this: Stop criticizing the status quo applications and think about where this is going. In mid-January, I said they'd add internet access and a "calculator." Both have already been added. OpenAI reduce the "hallucination" problem between ChatGPT3.5 and ChanTGPT4. On April 25, Nvidia claimed its new chips can prevent (or at least reduce) the problem by enabling other LLMs to confirm questionable facts.
Perplexity.ai -- Gets Better!
Perplexity.ai is made some substantial improvements, including the integration of ChatGPT4. If your school has blocked OpenAI/ChatGPT (and even if it hasn't), you should try perplexity.ai. It also doesn't require an individual log on, so there are fewer privacy protections. I need to spend some time on this at the last part of the week on this. Alexey Romanov .
Confidence
I'm a little surprised by how confident they are about issues like: Yes, it will destroy jobs; No, it won't cause labor disruptions; ues, AGI soon; no, no way we can have AGI. There are many unknowns and many very smart people saying different things. Maybe we should be a bit less confident? There are ranges, but things seem quite unpredictable, as we've never had this rate of technological advance and we've never lived among machines with degrees (depending on your take) of intelligence.
Deep Fakes
There is a lot new with Deep Fakes this week.
A new study reveals that humans cannot distinguish between real photos and AI-created fake photos to a significant degree 38.7%."
Someone "animated" a Delta commercial using AI art tools. You have to look very closely to realize the people aren't real.
AI is Dangerous without Blockchain Endrit Restelica . This is something I need to start to understand/unpack.
There needs to be a serious discussion of this in schools. And outside of schools -- As Lee Tiedrich notes: We need a global AI learning campaign.
Thinking of Some Different Ways to Use ChatGPT in Education
General Benefits
How Close are We Meaningful AI Tutors in Education?
Application: Checks for Understanding. Mike Sharples used ChatGPT to tutor him about quantum computing assess his knowledge along the way.
Are You a K-6 Teacher Looking for Practical Uses?
Teachers Need to Start Using ChatGPT/Generative AI
I found Bonnie Nieves ' How Teachers Model a Growth Mindset through AI Integration. "By embracing AI, educators demonstrate their commitment to lifelong learning, adaptability, and resilience, inspiring students to adopt a similar mindset and succeed in an ever-evolving world."
Ethan Mollick makes some important points here. I'll highlight three:
(1) Teachers should experiment with ChatGPT and learn how to use it. It will help them and their students a lot and it's more useful than learning a ton of new apps.
(2) We can't really stop plagiarism and the technology improves every day (can do math, fewer problems with facts, humans often get facts wrong).
(3) EdTech companies may be in trouble because ChatGPT can do what they can do for a fraction of the price (even $0). He's probably underestimating the power of lobbies and the "cards" companies will play to say that only their ChatGPT skin guarantees a safe user experience and complies with privacy law. Lobbyistss. and snake oil salesmen are good at what they do.
Student Feedback
Participate in Dr. Nick Jackson's opportunities for student feedback this summer!
Web Extenders!
I've been ranting about this for three months -- ChatGPT doesn't search the internet when you enter a prompt, so you can't use it for things like bibliographies. But, you can use the Chrome web search extender, the Bing integration and perplexity.ai (more on that later).
Recommended by LinkedIn
What I didn't know was that you can restrict your search to references from Google Scholar and others!
Writing
Grading
College Writing
AI in Academic Scholarship
Ethics -- General
AI has social consequences, but who pays the price? Tech companies’ problem with ‘ethical debt’. There are unintended consequences from AI advances and we won't know what they all are until the we hit some of the consequences.
Ethics -- Language Extinction
Ethics -- Art
Ethics -- Unemployment
I visited a stealth startup where every employee is using ChatGPT to do 10+ remote work jobs at other companies. Big tech, AI, fintech, crypto, you name it.
They already control ~15% of all tech jobs at all these other companies.
They are going to be opening more warehouses in the future, and continuing to scale up efforts, until they control *all tech jobs*
Unemployment -- Teachers
Teachers worry about losing their jobs to AI, but we don't have enough teachers, and this problem isn't limited to the US. Barbara Anna Zielonka
There has been a sharp decline in the number of applicants to teacher education programs in Norway, according to figures from The Norwegian Universities and Colleges Admission Service (Samordna opptak). There are 21.9 percent fewer applicants for teacher education this year than last year. I am not surprised by these numbers, but we really need to focus on how to retain teachers/ increase teacher retention rates NOW!
Ethics -- Universal Basic Income
Ethics -- Health Care
Education Needs Federal Funding for AI Adaption!
Artificial General Intelligence
These technologies will be able to do a lot before they are on, average, as smart as us and possess all our intelligence capabilities, but I've found that the details of the debate can help you learn about where we are headed.
Dr. Jennifer Prendki had a great post on Sunday about mental models that was really interesting and understandable for lay folks like me. It also drew some great comments about things I hadn't thought of. Someone also linked this very awesome lecture. My bottom line is that there is some very amazing debate on this issue by some truly brilliant people but that nearly everyone agrees that we will obtain AGI and that the timeline for that is getting faster and faster and that along the way the machines will match many of our capabilities.
Learn more about the brain and parts of the brain current AI capabilities are developing. in:
End of the World?
Big Moves
Google merges its Deep Mind and Google Brain teams in an effort to accelerate its AI developments and to compete with Microsoft and ChatGPT.