Remember Microsoft’s shiny new AI tool, “Recall”? It’s like your personal time machine, answering questions about your browsing history and laptop activity by taking screenshots every 5 seconds. Sounds cool, right? Well, it gets problematic.
What’s going on here?
Security researchers have found a potential privacy nightmare lurking within this seemingly convenient tool.
What does this mean?
Recall stores all those screenshots in an unencrypted database on your laptop. This means anyone with access to your device could potentially see everything you’ve been doing. Cybersecurity experts are already comparing it to spyware, and one ethical hacker even built a tool called “TotalRecall” (yes, like the movie) that can pull all the information Recall saves. Yikes.
Microsoft is partnering with Khan Academy in a multifaceted deal to demonstrate how AI can transform the way we learn. The cornerstone of today’s announcement centers on Khan Academy’s Khanmigo AI agent. Microsoft says it will migrate the bot to its Azure OpenAI Service, enabling the nonprofit educational organization to provide all U.S. K-12 educators free access to Khanmigo.
In addition, Microsoft plans to use its Phi-3 model to help Khan Academy improve math tutoring and collaborate to generate more high-quality learning content while making more courses available within Microsoft Copilot and Microsoft Teams for Education.
One in three American teachers have used artificial intelligence tools in their teaching at least once, with English and social studies teachers leading the way, according to a RAND Corporation survey released last month. While the new technology isn’t yet transforming how kids learn, both teachers and district leaders expect that it will become an increasingly common feature of school life.
When ChatGPT emerged a year and half ago, many professors immediately worried that their students would use it as a substitute for doing their own written assignments — that they’d click a button on a chatbot instead of doing the thinking involved in responding to an essay prompt themselves.
But two English professors at Carnegie Mellon University had a different first reaction: They saw in this new technology a way to show students how to improve their writing skills.
“They start really polishing way too early,” Kaufer says. “And so what we’re trying to do is with AI, now you have a tool to rapidly prototype your language when you are prototyping the quality of your thinking.”
He says the concept is based on writing research from the 1980s that shows that experienced writers spend about 80 percent of their early writing time thinking about whole-text plans and organization and not about sentences.
On Building AI Models for Education — from aieducation.substack.com by Claire Zau Google’s LearnLM, Khan Academy/MSFT’s Phi-3 Models, and OpenAI’s ChatGPT Edu
This piece primarily breaks down how Google’s LearnLM was built, and takes a quick look at Microsoft/Khan Academy’s Phi-3 and OpenAI’s ChatGPT Edu as alternative approaches to building an “education model” (not necessarily a new model in the latter case, but we’ll explain). Thanks to the public release of their 86-page research paper, we have the most comprehensive view into LearnLM. Our understanding of Microsoft/Khan Academy small language models and ChatGPT Edu is limited to the information provided through announcements, leaving us with less “under the hood” visibility into their development.
Answer AI is among a handful of popular apps that are leveraging the advent of ChatGPT and other large language models to help students with everything from writing history papers to solving physics problems. Of the top 20 education apps in the U.S. App Store, five are AI agents that help students with their school assignments, including Answer AI, according to data from Data.ai on May 21.
If your school (district) or university has not yet made significant efforts to think about how you will prepare your students for a World of AI, I suggest the following steps:
July 24 – Administrator PD & AI Guidance In July, administrators should receive professional development on AI, if they haven’t already. This should include…
August 24 –Professional Development for Teachers and Staff…
Fall 24 — Parents; Co-curricular; Classroom experiments…
December 24 — Revision to Policy…
New ChatGPT Version Aiming at Higher Ed — from insidehighered.com by Lauren Coffey ChatGPT Edu, emerging after initial partnerships with several universities, is prompting both cautious optimism and worries.
OpenAI unveiled a new version of ChatGPT focused on universities on Thursday, building on work with a handful of higher education institutions that partnered with the tech giant.
The ChatGPT Edu product, expected to start rolling out this summer, is a platform for institutions intended to give students free access. OpenAI said the artificial intelligence (AI) toolset could be used for an array of education applications, including tutoring, writing grant applications and reviewing résumés.
Introducing ChatGPT Edu — from openai.com An affordable offering for universities to responsibly bring AI to campus.
We’re announcing ChatGPT Edu, a version of ChatGPT built for universities to responsibly deploy AI to students, faculty, researchers, and campus operations. Powered by GPT-4o, ChatGPT Edu can reason across text and vision and use advanced tools such as data analysis. This new offering includes enterprise-level security and controls and is affordable for educational institutions.
We built ChatGPT Edu because we saw the success universities like the University of Oxford, Wharton School of the University of Pennsylvania(opens in a new window), University of Texas at Austin, Arizona State University(opens in a new window), and Columbia University in the City of New York were having with ChatGPT Enterprise.
ChatGPT can help with various tasks across campus, such as providing personalized tutoring for students and reviewing their resumes, helping researchers write grant applications, and assisting faculty with grading and feedback.
Tool use, which enables Claude to interact with external tools and APIs,is now generally available across the entire Claude 3 model family on the Anthropic Messages API, Amazon Bedrock, and Google Cloud’s Vertex AI. With tool use, Claude can perform tasks, manipulate data, and provide more dynamic—and accurate—responses.
Define a toolset for Claude and specify your request in natural language. Claude will then select the appropriate tool to fulfill the task and, when appropriate, execute the corresponding action:
Extract structured data from unstructured text…
Convert natural language requests into structured API calls…
Answer questions by searching databases or using web APIs…
Automate simple tasks through software APIs…
Orchestrate multiple fast Claude subagents for granular tasks…
From DSC: The above posting reminds me of this other posting…as AGENTS are likely going to become much more popular and part of our repertoire:
Forget Chatbots. AI Agents Are the Future — from wired.com by Will Knight Startups and tech giants are trying to move from chatbots that offer help via text, to AI agents that can get stuff done. Recent demos include an AI coder called Devin and agents that play videogames.
Devin is just the latest, most polished example of a trend I’ve been tracking for a while—the emergence of AI agents that instead of just providing answers or advice about a problem presented by a human can take action to solve it. A few months back I test drove Auto-GPT, an open source program that attempts to do useful chores by taking actions on a person’s computer and on the web. Recently I tested another program called vimGPT to see how the visual skills of new AI models can help these agents browse the web more efficiently.
Copilot+ PCs are the fastest, most intelligent Windows PCs ever built. With powerful new silicon capable of an incredible 40+ TOPS (trillion operations per second), all–day battery life and access to the most advanced AI models, Copilot+ PCs will enable you to do things you can’t on any other PC. Easily find and remember what you have seen in your PC with Recall, generate and refine AI images in near real-time directly on the device using Cocreator, and bridge language barriers with Live Captions, translating audio from 40+ languages into English.
From DSC: As a first off-the-hip look, Recall could be fraught with possible security/privacy-related issues. But what do I know? The Neuron states “Microsoft assures that everything Recall sees remains private.” Ok…
From The Rundown AI concerning the above announcements:
The details:
A new system enables Copilot+ PCs to run AI workloads up to 20x faster and 100x more efficiently than traditional PCs.
Windows 11 has been rearchitected specifically for AI, integrating the Copilot assistant directly into the OS.
New AI experiences include a new feature called Recall, which allows users to search for anything they’ve seen on their screen with natural language.
Copilot’s new screen-sharing feature allows AI to watch, hear, and understand what a user is doing on their computer and answer questions in real-time.
Copilot+ PCs will start at $999, and ship with OpenAI’s latest GPT-4o models.
Why it matters: Tony Stark’s all-powerful JARVIS AI assistant is getting closer to reality every day. Once Copilot, ChatGPT, Project Astra, or anyone else can not only respond but start executing tasks autonomously, things will start getting really exciting — and likely initiate a whole new era of tech work.
AI’s New Conversation Skills Eyed for Education — from insidehighered.com by Lauren Coffey The latest ChatGPT’s more human-like verbal communication has professors pondering personalized learning, on-demand tutoring and more classroom applications.
ChatGPT’s newest version, GPT-4o ( the “o” standing for “omni,” meaning “all”), has a more realistic voice and quicker verbal response time, both aiming to sound more human. The version, which should be available to free ChatGPT users in coming weeks—a change also hailed by educators—allows people to interrupt it while it speaks, simulates more emotions with its voice and translates languages in real time. It also can understand instructions in text and images and has improved video capabilities.
…
Ajjan said she immediately thought the new vocal and video capabilities could allow GPT to serve as a personalized tutor. Personalized learning has been a focus for educators grappling with the looming enrollment cliff and for those pushing for student success.
There’s also the potential for role playing, according to Ajjan. She pointed to mock interviews students could do to prepare for job interviews, or, for example, using GPT to play the role of a buyer to help prepare students in an economics course.
A Guide to the GPT-4o ‘Omni’ Model — from aieducation.substack.com by Claire Zau The closest thing we have to “Her” and what it means for education / workforce
Today, OpenAI introduced its new flagship model, GPT-4o, that delivers more powerful capabilities and real-time voice interactions to its users. The letter “o” in GPT-4o stands for “Omni”, referring to its enhanced multimodal capabilities. While ChatGPT has long offered a voice mode, GPT-4o is a step change in allowing users to interact with an AI assistant that can reason across voice, text, and vision in real-time.
Facilitating interaction between humans and machines (with reduced latency) represents a “small step for machine, giant leap for machine-kind” moment.
Everyone gets access to GPT-4: “the special thing about GPT-4o is it brings GPT-4 level intelligence to everyone, including our free users”, said CTO Mira Murati. Free users will also get access to custom GPTs in the GPT store, Vision and Code Interpreter. ChatGPT Plus and Team users will be able to start using GPT-4o’s text and image capabilities now
ChatGPT launched a desktop macOS app: it’s designed to integrate seamlessly into anything a user is doing on their keyboard. A PC Windows version is also in the works (notable that a Mac version is being released first given the $10B Microsoft relationship)
In a surprise launch, OpenAI dropped GPT-4 Omni, their new leading model. They also made a bunch of paid features in ChatGPT free and announced a new desktop app. Pete breaks down what you should know and what this says about AI.
Hello GPT-4o — from openai.com We’re announcing GPT-4o, our new flagship model that can reason across audio, vision, and text in real time.
GPT-4o (“o” for “omni”) is a step towards much more natural human-computer interaction—it accepts as input any combination of text, audio, image, and video and generates any combination of text, audio, and image outputs. It can respond to audio inputs in as little as 232 milliseconds, with an average of 320 milliseconds, which is similar to human response time in a conversation. It matches GPT-4 Turbo performance on text in English and code, with significant improvement on text in non-English languages, while also being much faster and 50% cheaper in the API. GPT-4o is especially better at vision and audio understanding compared to existing models.
Providing inflection, emotions, and a human-like voice
Understanding what the camera is looking at and integrating it into the AI’s responses
Providing customer service
With GPT-4o, we trained a single new model end-to-end across text, vision, and audio, meaning that all inputs and outputs are processed by the same neural network. Because GPT-4o is our first model combining all of these modalities, we are still just scratching the surface of exploring what the model can do and its limitations.
This demo is insane.
A student shares their iPad screen with the new ChatGPT + GPT-4o, and the AI speaks with them and helps them learn in *realtime*.
Imagine giving this to every student in the world.
The rapid rise of artificial intelligence appears to be taking a toll on the shares of online education companies Chegg and Coursera.
Both stocks sank by more than 10% on Tuesday after issuing disappointing guidance in part because of students using AI tools such as ChatGPT from OpenAI.
“Combining AI Literacy with Core Humanities Learning Goals: Practical, Critical, and Playful Approaches”
It was amazing to get to do an in-person keynote at @csunorthridge‘s AI Pedagogy Showcase.
Sharing slides: https://t.co/LDIGAZ3ORO 1/2
— Anna Mills, annamillsoer.bsky.social, she/her (@AnnaRMills) May 3, 2024
Synthetic Video & AI Professors— from drphilippahardman.substack.com by Dr. Philippa Hardman Are we witnessing the emergence of a new, post-AI model of async online learning?
TLDR: by effectively tailoring the learning experience to the learner’s comprehension levels and preferred learning modes, AI can enhance the overall learning experience, leading to increased “stickiness” and higher rates of performance in assessments.
…
TLDR: AI enables us to scale responsive, personalised “always on” feedback and support in a way that might help to solve one of the most wicked problems of online async learning – isolation and, as a result, disengagement.
…
In the last year we have also seen the rise of an unprecedented number of “always on” AI tutors, built to provide coaching and feedback how and when learners need it.
Perhaps the most well-known example is Khan Academy’s Khanmigo and its GPT sidekick Tutor Me. We’re also seeing similar tools emerge in K12 and Higher Ed where AI is being used to extend the support and feedback provided for students beyond the physical classroom.
Given the potential ramifications of artificial intelligence (AI) diffusion on matters of diversity, equity, inclusion, and accessibility, now is the time for higher education institutions to adopt culturally aware, analytical decision-making processes, policies, and practices around AI tools selection and use.
The theme of the day was Human Connection V Innovative Technology. I see this a lot at conferences, setting up the human connection (social) against the machine (AI). I think this is ALL wrong. It is, and has always been a dialectic, human connection (social) PLUS the machine. Everyone had a smartphone, most use it for work, comms and social media. The binary between human and tech has long disappeared.
I recently created an AI version of myself—REID AI—and recorded a Q&A to see how this digital twin might challenge me in new ways. The video avatar is generated by Hour One, its voice was created by Eleven Labs, and its persona—the way that REID AI formulates responses—is generated from a custom chatbot built on GPT-4 that was trained on my books, speeches, podcasts and other content that I’ve produced over the last few decades. I decided to interview it to test its capability and how closely its responses match—and test—my thinking. Then, REID AI asked me some questions on AI and technology. I thought I would hate this, but I’ve actually ended up finding the whole experience interesting and thought-provoking.
From DSC: This ability to ask questions of a digital twin is very interesting when you think about it in terms of “interviewing” a historical figure. I believe character.ai provides this kind of thing, but I haven’t used it much.
Since AI in education has been moving at the speed of light, we built this AI Tools in Education database to keep track of the most recent AI tools in education and the changes that are happening every day. This database is intended to be a community resource for educators, researchers, students, and other edtech specialists looking to stay up to date. This is a living document, so be sure to come back for regular updates.
The day started out with a short talk about AI (slides). Some of it is my usual schtick where I do a bit of Q&A with folks around myths and misunderstandings of generative AI in order to establish some common ground. These are often useful both in setting the tone and giving folks a sense of how I come to explore generative AI: with a mixture of humor, concern, curiosity, and of course, cat pics.
From there, we launched into a series of mini-workshops where folks had time to first play around with some previously created prompts around teaching and learning before moving onto prompts for administrative work. The prompts and other support materials are in this Workshop Resource Document. The goal was to just get them into using one or more AI tools with some useful prompts so they can learn more about its capabilities.
The Edtech Insiders Rundown of ASU+GSV 2024 — from edtechinsiders.substack.com by by Sarah Morin, Alex Sarlin, and Ben Kornell And more on Edtech Insiders+, upcoming events, Gauth, AI Reading Tutors, The Artificial Intelligence Interdisciplinary Institute, and TeachAI Policy Resources
Alex Sarlin
4. Everyone is Edtech Now This year, in addition to investors, entrepreneurs, educators, school leaders, university admins, non-profits, publishers, and operators from countless edtech startups and incumbents, there were some serious big tech companies in attendance like Meta, Google, OpenAI, Microsoft, Amazon, Tiktok, and Canva. Additionally, a horde of management consultancies, workforce organizations, mental health orgs, and filmmakers were in attendance.
Edtech continues to expand as an industry category and everyone is getting involved.
It was such a delight to chat with Ana. She’s brilliant and passionate, a talented educator, and an advocate for better ways of learning for children and adults. We cover ways to transform schools so that students get real-world skills, learn resilience and how to embrace challenges, and are prepared for an unpredictable future. And we go hard on why we must keep learning no matter our age, become generalists, and leverage technology in order to adapt to the fast-changing world.
The Texas Tribune reports an “automated scoring engine” that utilizes natural language processing — the technology that enables chatbots like OpenAI’s ChatGPT to understand and communicate with users — is being rolled out by the Texas Education Agency (TEA) to grade open-ended questions on the State of Texas Assessments of Academic Readiness (STAAR) exams. The agency is expecting the system to save $15–20 million per year by reducing the need for temporary human scorers, with plans to hire under 2,000 graders this year compared to the 6,000 required in 2023.
Debating About AI: An Easy Path to AI Awareness and Basic Literacy — from stefanbauschard.substack.com by Stefan Bauschard If you are an organization committed to AI literacy, consider sponsoring some debate topics and/or debates next year and expose thousands of students to AI literacy.
Resolved: Teachers should integrate generative AI in their teaching and learning.
The topic is simple but raises an issue that students can connect with.
While helping my students prepare and judging debates, I saw students demonstrate an understanding of many key issues and controversies.
These included—
*AI writing assessment/grading
*Bias
*Bullying
*Cognitive load
*Costs of AI systems
*Declining test scores
*Deep fakes
*Differentiation
*Energy consumption
*Hallucinations
*Human-to-human connection
*Inequality and inequity in access
*Neurodiversity
*Personalized learning
*Privacy
*Regulation (lack thereof)
*The future of work and unemployment
*Saving teachers time
*Soft skills
*Standardized testing
*Student engagement
*Teacher awareness and AI training; training resource trade-offs
*Teacher crowd-out
*Transparency and explainability
*Writing detectors (students had an exaggerated sense of the workability of these tools).
[Per Chris Dede] We often think of the product of teaching as the outcome (e.g. an essay, a drawing, etc.). The essence of education, in my view, lies not in the products or outcomes of learning but in the journey itself. The artifact is just a symbol that you’ve taken the journey.
The process of learning — the exploration, challenges, and personal growth that occur along the way — is where the real value lies. For instance, the act of writing an essay is valuable not merely for the final product but for the intellectual journey it represents. It forces you to improve and organize your thinking on a subject.
This distinction becomes important with the rise of generative AI, because it uniquely allows us to produce these artifacts without taking the journey.
As I’ve argued previously, I am worried that all this hype around LLMs renders them a “type of digital duct-tape to hold together an obsolete industrial-era educational system”.
Speaking of AI in our learning ecosystems, also see:
Usage Guidelines for AI Generative Tools at College Unbound These guidelines were created and reviewed by College Unbound students in Spring 2023 with the support of Lance Eaton, Director of Faculty Development & Innovation. The students include S. Fast, K. Linder-Bey, Veronica Machado, Erica Maddox, Suleima L., Lora Roy.
A recent study has found that scientific citations generated by ChatGPT often do not correspond to real academic work. The study, published in the Canadian Psychological Association’s Mind Pad, found that “false citation rates” across various psychology subfields ranged from 6% to 60%. Surprisingly, these fabricated citations feature elements such as legitimate researchers’ names and properly formatted digital object identifiers (DOIs), which could easily mislead both students and researchers.
…
MacDonald found that a total of 32.3% of the 300 citations generated by ChatGPT were hallucinated. Despite being fabricated, these hallucinated citations were constructed with elements that appeared legitimate — such as real authors who are recognized in their respective fields, properly formatted DOIs, and references to legitimate peer-reviewed journals.
The artificial intelligence sector has never been more competitive. Forbes received some 1,900 submissions this year, more than double last year’s count. Applicants do not pay a fee to be considered and are judged for their business promise and technical usage of AI through a quantitative algorithm and qualitative judging panels. Companies are encouraged to share data on diversity, and our list aims to promote a more equitable startup ecosystem. But disparities remain sharp in the industry. Only 12 companies have women cofounders, five of whom serve as CEO, the same count as last year. For more, see our full package of coverage, including a detailed explanation of the list methodology, videos and analyses on trends in AI.
New Generative AI video tools coming to Premiere Pro this year will streamline workflows and unlock new creative possibilities, from extending a shot to adding or removing objects in a scene
Adobe is developing a video model for Firefly, which will power video and audio editing workflows in Premiere Pro and enable anyone to create and ideate
Adobe previews early explorations of bringing third-party generative AI models from OpenAI, Pika Labs and Runway directly into Premiere Pro, making it easy for customers to draw on the strengths of different models within the powerful workflows they use every day
AI-powered audio workflows in Premiere Pro are now generally available, making audio editing faster, easier and more intuitive
