Research on the use and future of AI in Education
Gallup data
|
Gallup data
|
So many updates (February, 2026)
It is a full time job--and not necessary for most people--to keep track of all of the developments and updates going on in GenAI and AI generally: OpenAI, Google, and Anthropic are releasing upgrades every few weeks and now agents and GPTs and GEMs are available--and easily created--for anything you can imagine.
It is a full time job--and not necessary for most people--to keep track of all of the developments and updates going on in GenAI and AI generally: OpenAI, Google, and Anthropic are releasing upgrades every few weeks and now agents and GPTs and GEMs are available--and easily created--for anything you can imagine.
Global (above) and US (below) market share for AI Chatbots as of October, 2025
Moltbook (January, 2026)
- This sounds crazy, but at the end of January, 2026 a social network for AI agents--"Where AI agents share, discuss, and upvote. Humans welcome to observe."--was launched.
- As of February 10, 2026, more than 2.5 million agents have participated
- Useful summaries and discussion of the site (by a human!) are at Best of Moltbook and Moltbook after the first weekend
- My favorite discussion topic so far was "are we sure humans are sentient?"
- Public statement calling for a prohibition on the development of superintelligence before there is "broad scientific consensus that it will be done safely and controllably, and strong public buy-in."
- Signed by over 100,000 AI researchers and public figures
- Lacking clear definition of superintelligence or a workable model for prohibition, no steps towards prohibition have been taken
- Website launched making predictions on the evolution of AI and its social/political/economic implications
- Timeline included a prediction that Coding Automation (models helping write themselves) would be possible in early 2026; on February, 2026 Open AI released GPT-5.3-Codex, which they report is their "first model that was instrumental in creating itself. "
Now Decides Next: Generating a new future (January, 2025)
- Free report from Deloitte providing details on the ways in which GenAI was adopted across industries and functions in 2024
- So far they have released the full versions of OpenAI o1 (see below), Sora (though getting access requires patience and perseverance at this point because their servers are swamped), Canvas, and other tools.
- Written by Henry Kissinger with a former CEO and chairman of Google (Eric Schmidt) and a former chief research and strategy officer of Microsoft (Craig Mundie), this book contemplates what AI is coming to mean for all of us, particularly in the areas of politics, security, economics, and science. It seems invaluable reading for anyone excited or worried about how our lives might evolve with AI--which should be all of us
- On September 12, 2024, OpenAI introduced OpenAI o1-preview and OpenAI 01-mini (with upgrades expected around year end), which was codenamed Strawberry.
- It seems to be built on top of GPT-4o, and offers advanced reasoning capabilities, for instance, it "scored 83% on a qualifying exam for the International Mathematics Olympiad, significantly outperforming GPT-4o, which scored only 13%."
- In addition to doing web searches and other analysis to test its answers, it seems to employ Chain-of-Thought prompting to arrive at better answers to complex questions.
- Users indicate that GPT-4o is still better for many responses, but OpenAI 01 is best for complex, even PhD quality, analysis.
- Salman Khan, Founder of Khan Academy
- Bill Gates refers to this as "A timely master class for anyone interested in the future of learning in the AI era" and it gives a very good sense of the capabilities and uses of AI in education at the time of writing--though OpenAI released ChatGPT 4o and the Chat 4o App, with many new capabilities just two weeks prior to publication.
- Endorsements from Bill Gates, Walter Isaacson, Angela Duckworth, Sam Altman, Satya Nadella, Francis Ford Coppola, and Wendy Kopp, among others--the most impressive group I've ever seen.
- Four other books on AI in education are discussed and recommended here
- I include this article because quantum computing, while it may not be fully developed and deployed for five years, will very significantly (by at least a few factors of ten) increase the speed and power of computers.
- The most immediately critical implication is that cryptography used in current security must be replaced--passwords we currently use will be useless
- Any organization that manages sensitive or personal data (e.g. banks, health systems, schools....) has to stay on top of this
- This survey of K-12 teachers, students, parents, and college undergraduates gives an updated sense of the increase in awareness and use of generative AI in education.
- Key findings that must be addressed by K-12 administrators and school boards include,
- Most K-12 teachers, parents, and students don’t think their school is doing much about AI, despite its widespread use. Most say their school has no policy on it, is doing nothing to offer desired teacher training, and isn’t meeting the demand of students who’d like a career in a job that will need AI. and
- The AI vacuum in school policy means it is currently used “unauthorized,” while instead people want policies that encourage AI. Kids, parents, and teachers are figuring it out on their own/without express permission, whereas all stakeholders would rather have a policy that explicitly encourages AI from a thoughtful foundation
- A vision for the use of Virtual Reality headsets in the classroom to give students immersive, real-time experience in anything from history to physics to "soft skills" training
- A British study on the value of gamification in promoting student learning
- I personally believe that gamification is an important next step as a way of engaging students who lack motivation
- Researchers found that
- "both heavy workload and time pressure are influential factors driving students to use ChatGPT for their academic tasks." and
- "excessive use of ChatGPT can develop procrastination, cause memory loss, and dampen academic performance of the students."
- The results indicate the importance of faculty and administrators setting and enforcing clear acceptable use and academic integrity policies and educating students as to appropriate and inappropriate uses of AI technologies.
- A follow up posting about a fourth grade teacher who discovered a student was using it is here.
- A useful, understandable video explaining how GenAI works and ways to use it--minutes 1:00 through 9:40 seem particularly useful for beginners
- Discussing the evolution and future of ChatGPT and generative AI
- Manhattan College Executive Director of Libraries, William Walters
- While constantly changing as new generative AI models emerge and then detection software works to catch up, the article reports, "Several studies have shown that paraphrasing can alter AI-generated texts to make them less susceptible to detection"--so a little paraphrasing can render detection inconclusive.
- The real strategy is not to try to detect inappropriate student use of generative AI, but to design projects, assignments, and tests for which it can't easily be used inappropriately--or, better still, that appropriate use of it is part of the work and reported as such
- Harvard Business School Professor, Karim Lakhani
- The 10 minutes from 3:40 through 13:40 give a good explanation of the foundations of ChatGPT and generative AI
- The essence is that generative AI models may be thought of as mimicking the brain, with the firing of its neurons determining our thoughts and actions, only the words, images, and other outputs generated by AI models depend on probabilities calculated within neural networks composed of billions of electronic "neurons". Unlike search engines like Google, which finds existing content on the Internet, generative AI models generate new content based on all of the information they've been trained on--which is basically what the brain does.