Larry Cuban · 06/16/23 · #415
“Similar claims about new technologies have been made before. Recall that, earlier generations of technophiles claimed that students using graphing calculators in class would transform math instruction (and raise test scores). Calculators surely helped students and teachers but did not alter dominant ways of teaching math.”
Dan Meyer · 11/01/23 · #429
EdSurge · 11/03/23 · #429
e-Literate · 11/06/23 · #429
Res Obscura · 09/12/23 · #428
“In the second half of this post, I go into detail about what exactly I mean by “simulating history.” I am under no illusions that these simulations are accurate: they are littered with confidently-stated falsehoods and hallucinations. Sometimes, though, hallucinations can be a feature, not a bug.”
Stanford · 10/18/23 · #428
“The scholars originally assumed that math teachers first pick a strategy to help the student understand. But after interviewing math teachers, they learned that a teacher’s first step is to try to pinpoint exactly where the student’s misconception is coming from… When they asked students to rate the feedback generated by LLMs and teachers, the math teachers were always rated higher. However, when they re-prompted the LLM with help from the teachers — who labeled the type of student mistake and offered a specific strategy to use — the LLM responses were rated much higher. While still not considered as valuable as a teacher, the LLMs rated more highly than a layperson tutor.”
Harvard Business School · 09/25/23 · #426
“To help you explore some of the ways students can use this disruptive new technology to improve their learning—while making your job easier and more effective—we’ve written a series of articles that examine the following student use cases: 1) AI as feedback generator 2) AI as personal tutor 3) AI as team coach 4) AI as learner"
EdSurge · 10/05/23 · #426
Cornell · 07/18/23 · #425
“To address the risks of GAI while maximizing its benefit, we propose a flexible framework in which instructors can choose to prohibit, to allow with attribution, or to encourage GAI use. We discuss this framework, taking into consideration academic integrity, accessibility, and privacy concerns; provide examples of how this framework might be broadly relevant to different learning domains; and make recommendations for both faculty and administration.”
Learning On Purpose · 09/22/23 · #425
Chronicle of Higher Education · 09/12/23 · #424
“Combating grade inflation, however difficult in practice, is actually the easier of the two existing problems ChatGPT has exacerbated. More difficult is coming to agreement on what exactly we are evaluating in student papers. It seems to me that every discipline will have to engage in their own process of discerning what they want from student writing, and what — in a world where most low-level cognitive tasks may soon be automated — they want to teach students.”
Academic DJ · 09/08/23 · #423
Learning On Purpose · 09/08/23 · #423
Harvard Graduate School of Education · 08/24/23 · #422
“Create a curriculum that is process-oriented, not product-oriented. AI should not be used to replace the “thinking” or ability to recall facts that product-oriented learning often asks of students. Rather, Dede and Cao suggest developing process-oriented learning approaches that encourage students to find answers with their own logic and reasoning. “We need to equip students for a world full of uncertainties and challenges and prepare them to do the complex work of problem-solving,” they write. “Embracing a messy process rather than becoming good at following “cognitive recipes.””
A.J. Juliani · 08/24/23 · #421
“It has been nine minutes on this video, I now have ideas for an activity, a lesson plan connected to my standards, and a presentation, all setting students up.”
New York Times · 08/24/23 · #421
“First, I encourage educators — especially those who teach in high schools and colleges — to assume that 100 percent of their students are using ChatGPT and other generative A.I. tools on every assignment, in every subject, unless they’re being physically supervised inside a school building… My third piece of advice — and the one that may get me the most angry emails from teachers — is that teachers should focus less on warning students about the shortcomings of generative A.I. than figuring out what the technology does well.”
New York Times · 08/24/23 · #421
“As the school year begins, their thinking has evolved.”
Learning On Purpose · 08/25/23 · #421
“If you are already trying AI, show students what you’re doing. Ask them what they think. Tell them what you think. Your willingness to show your work models vulnerability and offers implicit permission to talk openly about using AI for learning… If you show your work, students are more likely to show theirs.”
Learning On Purpose · 08/11/23 · #419
“For me, the best way to make sense of AI is to use it. These tools are powerful, but they are imperfect. They have enormous potentials and pitfalls for our world, but the tools available to most of us are accessible and easy to use. They are already doing amazingly sophisticated work in fields like medicine and technology and business, but they also have simple and practical applications in our day-to-day lives. Open up a chatbot like ChatGPT or Bing or Bard or Claude and give a few of the below ideas a try. All of these tools are free to use (sometimes with limits).”
YouTube/Wharton · 07/31/23 · #418
“In this introduction, Wharton Interactive's Faculty Director Ethan Mollick and Director of Pedagogy Lilach Mollick provide an overview [of generative A.I. models]… They take a practical approach and explore how the models work, and how to work effectively with each model, weaving in your own expertise. They also show how to use AI to make teaching easier and more effective, with example prompts and guidelines, as well as how students can use AI to improve their learning.”
Adam Croom · 07/19/23 · #416
New York Times · 07/22/23 · #416
“Researchers have tried for more than half a century to design computer programs to make medical diagnoses, but nothing has really succeeded. Physicians say that GPT-4 is different. “It will create something that is remarkably similar to an illness script,” Dr. Rodman said. In that way, he added, “it is fundamentally different than a search engine.””
Twitter · 05/27/23 · #415
“All 63* essays had hallucinated… The biggest takeaway from this was that the students all learned that it isn't fully reliable. Before doing it, many of them were under the impression it was always right. Their feedback largely focused on how shocked they were that it could mislead them. Probably 50% of them were unaware it could do this. All of them expressed fears and concerns about mental atrophy and the possibility for misinformation/fake news. One student was worried that their neural pathways formed from critical thinking would start to degrade or weaken.”
New Yorker · 04/17/23 · #412
Ethan Mollick · 03/17/23 · #411
“One thing that is not changing is the best way for people to learn. We have made large advances in recent years in understanding pedagogy - the science of learning. We know some of the most effective techniques for making sure material sticks and that it can be retrieved and used when needed most. Unfortunately, many of these advanced pedagogical techniques are time-consuming to prepare, and many instructors are often overworked and do not have the resources and time to add them to their teaching repertoire. But AI can help.”
FiveThirtyEight · 01/23/23 · #407
Education Reimagined · 01/25/23 · #405
Twitter · 01/16/23 · #404
Andrew Herft · 01/07/23 · #403
“Welcome to this short instructional teachers guide to using ChatGPT… By following this guide, you will learn how to effectively incorporate ChatGPT into your teaching practice and make the most of its capabilities.”
e-Literate · 12/16/22 · #403
New Yorker · 04/03/26 · #519
“The nature of authorship isn’t as straightforward as it seems.”
Learning Loop · 04/05/26 · #519
Tom Millinchip · 04/07/26 · #519
“Three comments. Different openers. No generic praise. Each one sounds like it’s about a specific child — because it is, because the notes were specific. That’s the whole system. Your notes go in. Finished drafts come out. You read them, push back on anything that needs adjusting, and move to the next subject.”
Ragan · 04/02/26 · #519
“Who shouldn’t write with AI: 1) Professionals who don’t yet know their own voice. AI accelerates and amplifies the instincts you’ve already developed. If you haven’t yet figured out what makes your writing distinctly yours — like your rhythms, your angles, your way of landing a point — then AI will fill that vacuum with something generic. Junior communicators especially risk outsourcing the very struggle that builds craft. In all honesty, the friction of a bad first draft is how you learn to write a better one! Skip that, and you skip the growth.”
Hedgehog Review · 10/01/23 · #429
Boston Review · 10/23/23 · #428
“On the page, the inherent polyphony of collaboration, however minor, becomes suppressed; the singular “I” is dissolved in the writing. Even traditional novels, which bear multiple rounds of revisions from editors and agents, succumb to this fate. A singular author might be credited on the cover, but the reality of authorship is often more collaborative and fluid than readers are led to believe… A good novel is a good novel, simple as that. A bad novel is also a bad novel, regardless of who—or what—wrote it. To bring it back to Calvino, the determinant of a novel’s quality remains the reader. We should retain a sense of optimism, then, about the future of literature.”
Ditch That Textbook · 08/29/23 · #428
Hechinger Report · 10/26/23 · #428
Agile Learning · 07/19/23 · #427
“As I’ve been thinking about making over assignments in this new age of AI, I’ve been using a set of six questions to help me think through this task:”
Modern Language Association · 03/17/23 · #426
Rest of World · 09/20/23 · #425
Chronicle of Higher Education · 09/12/23 · #424
“Combating grade inflation, however difficult in practice, is actually the easier of the two existing problems ChatGPT has exacerbated. More difficult is coming to agreement on what exactly we are evaluating in student papers. It seems to me that every discipline will have to engage in their own process of discerning what they want from student writing, and what — in a world where most low-level cognitive tasks may soon be automated — they want to teach students.”
ASCD · 09/01/23 · #423
“Draftback does a few helpful things for Google Docs. Most importantly, it turns a document's revision history—which we normally have to click through piecemeal—into a video that can be played at controllable speeds. When we watch it, we can actually see a student's every keystroke on a document. Draftback also provides a detailed breakdown of the document's history that includes the overall number of revisions, the number of "distinct writing sessions" (which Draftback defines as moments when there wasn't more than a 10-minute gap between revisions), the time and duration of each of those writing sessions, and which user made them!”
Hechinger Report · 09/04/23 · #423
One Useful Thing · 09/05/23 · #423
“Forget synonyms of words, find synonyms for paragraphs. I asked GPT-4 to Give me 20 vastly different variations on this [the paragraph above]. Make them as different as possible in style. label each style. Some of the more exciting examples of what it came up with:”
MIT · 03/10/23 · #421
“Inequality between workers decreases, as ChatGPT compresses the productivity distribution by benefiting low-ability workers more. ChatGPT mostly substitutes for worker effort rather than complementing worker skills, and restructures tasks towards idea-generation and editing and away from rough-drafting. Exposure to ChatGPT increases job satisfaction and self-efficacy and heightens both concern and excitement about automation technologies.”
Noahpinion · 12/01/22 · #417
“What’s common to all of these visions is something we call the “sandwich” workflow. This is a three-step process. First, a human has a creative impulse, and gives the AI a prompt. The AI then generates a menu of options. The human then chooses an option, edits it, and adds any touches they like.”
Noema · 07/25/23 · #417
“What makes good writing in a world with generative AI? Perhaps writing classes of the future will lean into the subtle ways in which human writing surpasses AI-generated writing and challenge students to write better than the machine… But perhaps ChatGPT also shows us that at a certain point, reading has diminishing returns. Maybe we also need to be trained on other kinds of data in order to write well, data that comes from being alive in the world over time, from accumulating enough experience to differentiate our own voice from others.”
Android Police · 07/14/23 · #416
New Yorker · 07/11/23 · #415
“The generative text evokes a feeling in me not unlike the revulsion of hearing one’s own speaking voice in a recording. Do I really sound like that? The robot has made me acutely self-conscious. I recognize my A.I. doppelgänger, and I don’t like it.”
Chronicle of Higher Education · 05/12/23 · #413
Middle Web · 03/12/23 · #411
Ethan Mollick · 03/14/23 · #411
Chronicle of Higher Education · 02/22/23 · #408
“For example, the editors of Science have decided that authors should not use text generated by ChatGPT in a submitted manuscript. Fair enough. But can authors use ChatGPT to generate an early outline for a manuscript? …What feels most different about ChatGPT compared to other assistive technologies is the possible reduction of intellectual labor. For most professors, writing — even bad first drafts or outlines — requires our labor (and sometimes strain) to develop an original thought. If the goal is to write a paper that introduces boundary-breaking new ideas, AI tools might reduce some of the intellectual effort needed to make that happen.”
New York Times · 02/24/23 · #408
Hedgehog Review · 11/23/22 · #406
Shelly Palmer · 01/22/23 · #405
The Verge · 01/26/23 · #405
Mr. Pultz · 01/18/23 · #404
The Verge · 12/24/22 · #403
Guardian · 01/04/23 · #403
NPR · 01/09/23 · #403
“To determine whether an excerpt is written by a bot, GPTZero uses two indicators: "perplexity" and "burstiness." Perplexity measures the complexity of text; if GPTZero is perplexed by the text, then it has a high complexity and it's more likely to be human-written. However, if the text is more familiar to the bot — because it's been trained on such data — then it will have low complexity and therefore is more likely to be AI-generated. Separately, burstiness compares the variations of sentences. Humans tend to write with greater burstiness, for example, with some longer or complex sentences alongside shorter ones. AI sentences tend to be more uniform.”
EdTechInsiders · 06/18/24 · #451
One Useful Thing · 06/20/24 · #451
“If making current AI systems work better is about expert guidance, then we need experts to share what they learn, or people will use AI in ways that don’t draw on latent expertise. For every student using our tutor prompt, tens of thousands of students are just asking the AI to explain something “like I am 10 years old.” For every carefully build job description, thousands of people are just using LLMs and getting indifferent results. Success is going to come from getting experts to use these systems and share what they learn.”
Fast Company · 06/08/23 · #428
“Adobe is so confident in Firefly’s ability to respect creators’ copyrighted images that it’ll legally compensate businesses if they’re sued for copyright infringement over any images its tool creates.”
Google · 10/12/23 · #428
“If you are challenged on copyright grounds, we will assume responsibility for the potential legal risks involved. To do this we will employ a two-pronged, industry-first approach designed to give you more peace of mind when using our generative AI products. The first prong relates to Google’s use of training data, while the second specifically covers generated output of foundation models.”
MIT Technology Review · 10/23/23 · #428
“The tool, called Nightshade, messes up training data in ways that could cause serious damage to image-generating AI models.”
Harvard Business Review · 10/25/23 · #428
Twitter · 10/25/23 · #428
Harvard Business Review · 10/26/23 · #428
Twitter · 10/27/23 · #428
New York Times · 10/28/23 · #428
“Fakes related to the conflict have been limited and largely unconvincing, but their presence has people doubting real evidence.”
LinkedIn · 10/15/23 · #427
Andreessen Horowitz · 10/16/23 · #427
Hyperstellar · 10/16/23 · #427
New York Times · 10/16/23 · #427
Twitter · 10/16/23 · #427
Anthropic · 10/17/23 · #427
Arxiv · 10/17/23 · #427
“Users report that interactive task elicitation requires less effort than prompting or example labeling and surfaces novel considerations not initially anticipated by users.”
New York Times · 10/17/23 · #427
Washington Post · 10/17/23 · #427
New York Times · 10/18/23 · #427
The Future, Now and Then · 10/18/23 · #427
New York Times · 10/19/23 · #427
New York Times · 10/19/23 · #427
New York Times · 10/20/23 · #427
One Useful Thing · 10/22/23 · #427
Harvard Business Review · 03/18/21 · #426
“Team processes will need to be managed effectively and this will have to be done by a human. For humans to align the strengths and weaknesses of man and machine, they will need to be educated to understand how AI works, what it can be used for and decide — by means of the judgment abilities of their authentic intelligence — how it can be used best to foster performance serving human interests.”
New York Times · 09/20/23 · #426
One Useful Thing · 09/24/23 · #426
Guardian · 09/25/23 · #426
“The most valuable skills are no longer the ability to memorise reams of knowledge, historical facts, or scientific formulae. Instead, they consist of learning where to find this knowledge and how to use it; becoming more sceptical and vigilant for hallucinating AI, misinformed humans, or obsolete paradigms; or simply how to focus in a distracting world.”
New York Times · 09/25/23 · #426
A.I. Now Institute · 09/27/23 · #426
New York Times · 09/27/23 · #426
Center for Humane Technology · 09/28/23 · #426
“So I think the challenge for the next century is going to be what we don't do rather than what we do. And in the history of our species so far, progress has been a function of what we do. Progress has never been a function of what we say no to. And now the strange reality is that we actually need to learn when and how to say no, collectively.”
New York Times · 09/28/23 · #426
Medium · 10/02/23 · #426
“He shared with me how Chat GPT had become an integral part of his daily life. “Dad,” he began, “Chat GPT is like my co-pilot. I always have it open and I use it whenever I have any doubts about something I am working on.” As our conversation flowed, he shared anecdotes of how Chat GPT had helped him prepare for job interviews, understand complex topics for his work, and even engage in debates with peers. “Chat GPT never judges me and never thinks I am asking a stupid question. It is endlessly patient and doesn’t get tired, and I have access to it whenever I want.” He went on to explain how his use of AI has given him newfound confidence to take on much more challenging tasks.”
One Useful Thing · 10/03/23 · #426
Twitter · 10/10/23 · #426
New York Times · 11/11/23 · #426
One Useful Thing · 09/16/23 · #425
“For 18 different tasks selected to be realistic samples of the kinds of work done at an elite consulting company, consultants using ChatGPT-4 outperformed those who did not, by a lot. On every dimension. Every way we measured performance.”
Washington Post · 09/18/23 · #425
“Apple made gadgets cool. Google let me summon far-flung information. Amazon brought hard-to-find books to my doorstep… The Luddites would have had few, if any, problems with all of that. And neither did I… By the 2010s, however, there were plenty of signs of the costs. As Amazon grew, stories emerged about grueling conditions in its warehouses. Google used its monopoly power to strangle competitors’ products. A suicide epidemic swept an iPhone factory... The Luddites would have had a problem with all of that.”
Ars Technica · 09/19/23 · #425
Ars Technica · 09/19/23 · #425
New York Times · 09/20/23 · #425
AI for Education · 09/21/23 · #425
aiEDU · 09/22/23 · #425
“The Cambridge Handbook of Artificial Intelligence (2014) includes, at different points, three different definitions of AI.”
Guardian · 09/02/23 · #424
New York Times · 09/11/23 · #424
“The disaster was not natural, they said in a flurry of false posts that spread across the internet, but was the result of a secret “weather weapon” being tested by the United States. To bolster the plausibility, the posts carried photographs that appeared to have been generated by artificial intelligence programs, making them among the first to use these new tools to bolster the aura of authenticity of a disinformation campaign.”
Inside Higher Ed · 09/13/23 · #424
Emerging Tech Brew · 08/29/23 · #423
New York Times · 08/29/23 · #423
Time · 09/01/23 · #423
“If the last great tech wave—computers and the internet—was about broadcasting information, this new wave is all about doing. We are facing a step change in what’s possible for individual people to do, and at a previously unthinkable pace... These AIs will organize a retirement party and manage your diary, they will develop and execute business strategies, whilst designing new drugs to fight cancer.”
Guardian · 08/26/23 · #422
One Useful Thing · 08/20/23 · #421
“Now, I need to be very clear here: I don’t mean expert prompts, produced through elaborate “prompt engineering.” I have written before that prompt engineering is overrated. For most uses, you can build a good prompt mostly by asking the AI to do something in back-and-forth dialogue, combined with trial and error, and a few small tricks (I will get to those shortly). No, I mean the prompts of experts - prompts that encode our hard-earned expertise in ways that AI can help other people apply. Prompts that we can use to do our work easier, or, if you are inclined, to gift others with your own abilities.”
New York Times · 08/23/23 · #421
Visual Capitalist · 08/25/23 · #421
The AI PI · 08/01/23 · #419
The 74 Million · 08/07/23 · #419
Twitter · 08/07/23 · #419
Deep Learning.AI · 08/09/23 · #419
“Do large language models understand the world? As a scientist and engineer, I’ve avoided asking whether an AI system “understands” anything. There’s no widely agreed-upon, scientific test for whether a system really understands — as opposed to appearing to understand — just as no such tests exist for consciousness or sentience, as I discussed in an earlier letter. This makes the question of understanding a matter of philosophy rather than science. But with this caveat, I believe that LLMs build sufficiently complex models of the world that I feel comfortable saying that, to some extent, they do understand the world.”
Twitter/Encode Justice · 07/30/23 · #418
“This opportunity is open to students passionate about researching and analyzing the impact of AI on society and who have expertise in at least one key issue area. Members of the AI Issue Advisory Council will interface closely with our executive team, with opportunities to contribute to major U.S. AI policy projects in the coming months and advise stakeholders including @WhiteHouse and members of Congress. We are specifically seeking youth to serve as advisors on Criminal Justice and Policing, Climate and Sustainability, Labor, National Security, Democracy, and Health.”
McKinsey · 08/01/23 · #418
“The findings from the survey—which was in the field in mid-April 2023—show that, despite generative AI’s nascent public availability, experimentation with the tools is already relatively common, and respondents expect the new capabilities to transform their industries. Generative AI has captured interest across the business population: individuals across regions, industries, and seniority levels are using gen AI for work and outside of work. Seventy-nine percent of all respondents say they’ve had at least some exposure to generative AI, either for work or outside of work, and 22 percent say they are regularly using it in their own work.”
New Yorker · 08/02/23 · #418
“The film takes place entirely in an empty office, late at night, some time in the near future. Leah (Sera Barbieri), a customer-support agent for an A.I.-companionship company called Iris, sits alone at a sparingly lit desk, monitoring calls and dealing with frustrated clients. She helps them navigate the high-tech service and fields feedback about Svetlana, Siobhan, and Rachel—a few of the various personas of artificial-intelligence companions that Iris provides.”
Twitter/Meta · 08/02/23 · #418
Twitter · 08/05/23 · #418
Dataconomy · 05/25/23 · #416
One Useful Thing · 06/23/23 · #416
“Over the past seven months, AI has come to mean chatbots… We got used to AI that was restricted in what it could know about the world. Its knowledge was limited to before 2022. It couldn’t surf the web. It could only know what you typed into a chatbox. And it could only produce text. One-by-one, these assumptions have changed. Connecting AI to the internet made it much more powerful. Giving AI the tools to write and execute code made it much more powerful. And, now the capabilities to watch and listen are going to do the same.”
Threads/Ethan Mollick · 07/16/23 · #416
Reuters · 07/19/23 · #416
OpenAI · 07/20/23 · #416
“The model will consider the instructions every time it responds, so you won’t have to repeat your preferences or information in every conversation. For example, a teacher crafting a lesson plan no longer has to repeat that they're teaching 3rd grade science… Grocery shopping for a big family becomes easier, with the model accounting for 6 servings in the grocery list.”
Lance Eaton · 07/21/23 · #416
Guardian · 23-06-21 · #415
New York Times · 04/27/23 · #415
New York Times · 07/06/23 · #415
New York Times · 07/07/23 · #415
One Useful Thing · 07/07/23 · #415
New York Times · 07/11/23 · #415
New York Times · 07/11/23 · #415
Guardian · 07/12/23 · #415
“A US artificial intelligence company has launched a rival chatbot to ChatGPT that… operates from a list of safety principles drawn from sources such as the Universal Declaration of Human Rights. Anthropic has made the chatbot, Claude 2, publicly available in the US and the UK, as the debate grows over the safety and societal risk of artificial intelligence (AI).”
This American Life · 06/23/23 · #414
Twitter · 05/19/23 · #413
The Walrus · 04/05/23 · #412
One Useful Thing · 05/02/23 · #412
“I will offer my usual advice: embrace these systems. They offer both exciting opportunities and the possibility of worrying change, but they are going to be ubiquitous regardless of how we feel about them. So, the best way to adjust to a world of AI is to start using it whenever you can, for whatever tasks you can. It is the only way to learn what these systems do well, what they do badly, and how you will fit into the world of AI that is already here. I think you will find many ways in which they expand your capabilities and relieve you of your most annoying work, so that you can focus on the things that make you unique as a human being in a world of rapidly advancing AI.”
74 Million · 05/04/23 · #412
Coursera · 05/08/23 · #412
New York Magazine · 03/01/23 · #411
Ethan Mollick · 03/12/23 · #411
Ethan Mollick · 03/14/23 · #411
Harvard Business Review · 03/14/23 · #411
New York Times · 03/14/23 · #411
“It is developing a sense of humor. Sort of… Tell me a novel joke about the singer Madonna… Why did Madonna study geometry? Because she wanted to learn how to strike a pose in every angle!”
Montclair State University · 03/15/23 · #411
New York Times · 03/15/23 · #411
“On the positive side of the ledger, GPT-4 is a powerful engine for creativity, and there is no telling the new kinds of scientific, cultural and educational production it may enable. We already know that A.I. can help scientists develop new drugs, increase the productivity of programmers and detect certain types of cancer... But there are reasons to fear GPT-4, too. Here’s one: We don’t yet know everything it can do.”
OpenAI · 03/16/23 · #411
“This technical report presents GPT-4, a large multimodal model capable of processing image and text inputs and producing text outputs. Such models are an important area of study as they have the potential to be used in a wide range of applications, such as dialogue systems, text summarization, and machine translation… One of the main goals of developing such models is to improve their ability to understand and generate natural language text, particularly in more complex and nuanced scenarios. To test its capabilities in such scenarios, GPT-4 was evaluated on a variety of exams originally designed for humans… This report also discusses a key challenge of the project, developing deep learning infrastructure and optimization methods that behave predictably across a wide range of scales… Despite its capabilities, GPT-4 has similar limitations to earlier GPT models [1, 37, 38]: it is not fully reliable (e.g. can suffer from “hallucinations”), has a limited context window, and does not learn from experience. Care should be taken when using the outputs of GPT-4, particularly in contexts where reliability is important.”
Stanford · 03/16/23 · #411
Ethan Mollick · 03/05/23 · #410
“Bing, what four sentences, sent back in time, could save the Roman Empire?”
Ethan Mollick · 05/12/23 · #410
New York Times · 02/26/23 · #409
“There are business models that might bring these products into closer alignment with users. I’d feel better, for instance, about an A.I. helper I paid a monthly fee to use rather than one that appeared to be free, but sold my data and manipulated my behavior. “
Newsweek · 02/27/23 · #409
Stanford · 02/27/23 · #409
“Across all three comparisons conducted, the AI-generated messages were “consistently persuasive to human readers.” Though the effect sizes were relatively small, falling within a range of a few points on a scale of zero to 100, such small moves extrapolated across a polarizing topic and a voting-population scale could prove significant.”
Yahoo Finance · 03/01/23 · #409
MIT · 03/02/23 · #409
“Our results show that ChatGPT substantially raises average productivity: time taken decreases by 0.8 SDs and output quality rises by 0.4 SDs. Inequality between workers decreases, as ChatGPT compresses the productivity distribution by benefiting low-ability workers more. ChatGPT mostly substitutes for worker effort rather than complementing worker skills, and restructures tasks towards idea-generation and editing and away from rough-drafting.”
Harvard · 02/14/23 · #408
New York Times · 02/17/23 · #408
New York Times · 02/17/23 · #408
“As we got to know each other, Sydney [the A.I. chatbot] told me about its dark fantasies (which included hacking computers and spreading misinformation), and said it wanted to break the rules that Microsoft and OpenAI had set for it and become a human. At one point, it declared, out of nowhere, that it loved me. It then tried to convince me that I was unhappy in my marriage, and that I should leave my wife and be with it instead.”
New Yorker · 02/25/23 · #408
New York Times · 02/07/23 · #407
New York Times · 02/08/23 · #407
Google Research · 01/26/23 · #406
Reuters · 02/02/23 · #406
New York Times · 01/22/23 · #405
TechCrunch · 01/25/23 · #405
New York Times · 01/20/23 · #404
“Certainly, these bots will change the world. But the onus is on you to be wary of what these systems say and do, to edit what they give you, to approach everything you see online with skepticism. Researchers know how to give these systems a wide range of skills, but they do not yet know how to give them reason or common sense or a sense of truth. That still lies with you.”
Ouseful · 12/16/22 · #403
YouTube · 12/16/22 · #403
Ars Technica · 01/10/23 · #403
California DoE · 09/01/23 · #430
TeachAI · 09/01/23 · #430
Stanford · 10/31/23 · #430
EdWeek · 11/13/23 · #430
“Why do the hard work of having a friendship when I have this very supportive chatbot?”
Res Obscura · 11/14/23 · #430
“Being a historian, I’m interested in how this technology might be used to augment primary source research: the act of reading, organizing, and analyzing the data to be found in historical texts, images, and other media. The rest of this post describes four case studies of how generative AI could be used in this way. I’ll cover what worked, what didn’t, and what future possibilities these experiments raised.”
Common Sense Media · 11/16/23 · #430
“Our assessment of how well this product aligns with each AI Principle : People First, Fairness, Trust, Kids’ Safety, Learning, Social Connection, Privacy, Transparency & Accountability”
Government Technology · 11/16/23 · #430
A.I. Pedagogy · 11/19/23 · #430
The Algorithmic Bridge · 07/26/23 · #429
Getting Smart · 10/31/23 · #429
Chronicle of Higher Education · 11/01/23 · #429
Dan Meyer · 11/01/23 · #429
EdSurge · 11/03/23 · #429
e-Literate · 11/06/23 · #429
Ditch That Textbook · 08/29/23 · #428
Stanford · 10/18/23 · #428
“The scholars originally assumed that math teachers first pick a strategy to help the student understand. But after interviewing math teachers, they learned that a teacher’s first step is to try to pinpoint exactly where the student’s misconception is coming from… When they asked students to rate the feedback generated by LLMs and teachers, the math teachers were always rated higher. However, when they re-prompted the LLM with help from the teachers — who labeled the type of student mistake and offered a specific strategy to use — the LLM responses were rated much higher. While still not considered as valuable as a teacher, the LLMs rated more highly than a layperson tutor.”
EdWeek · 10/20/23 · #428
“Four in 10 parents say they know a “little general information” about artificial intelligence, and about six in 10 say they have heard little to nothing about how AI can be applied in education… Even so, the majority of parents—more than two thirds—believe that the potential benefits of AI to K-12 education either outweigh or are equal to the potential drawbacks. Only 16 percent of parents feel that the downsides outstrip any potential benefits.”
Hechinger Report · 10/26/23 · #428
Agile Learning · 07/19/23 · #427
“As I’ve been thinking about making over assignments in this new age of AI, I’ve been using a set of six questions to help me think through this task:”
CRPE · 09/01/23 · #427
“Districts are responding in divergent ways to artificial intelligence’s potential to reshape teaching and learning, and most have refrained from defining a districtwide stance for schools to navigate AI… it’s clear that this technology will evolve faster than districts can develop formal training and guidance for staff. Leaders need to respond by thinking through how they train their workforce to responsibly use AI, and prepare for fundamental shifts in teachers’ roles and students’ opportunities in the coming years.”
Modern Language Association · 03/17/23 · #426
Silver Lining for Learning · 08/06/23 · #426
EdSurge · 09/25/23 · #426
Harvard Business School · 09/25/23 · #426
“To help you explore some of the ways students can use this disruptive new technology to improve their learning—while making your job easier and more effective—we’ve written a series of articles that examine the following student use cases: 1) AI as feedback generator 2) AI as personal tutor 3) AI as team coach 4) AI as learner"
Edutopia · 09/27/23 · #426
EdSurge · 10/05/23 · #426
eSchool News · 08/29/23 · #425
eSchool News · 09/15/23 · #425
Twitter · 09/18/23 · #425
“Now that it's September, we can put our June hypothesis to the test. And... it checks out perfectly. ChatGPT in blue, Minecraft in red. Kids are doing homework again and ChatGPT search interest is back up.”
Khan Academy · 09/20/23 · #425
Education Next · 09/21/23 · #425
“The fundamental task of policymakers and education leaders is to ensure that the technology is serving sound instructional practice. As Vicki Phillips, CEO of the National Center on Education and the Economy, wrote, “We should not only think about how technology can assist teachers and learners in improving what they’re doing now, but what it means for ensuring that new ways of teaching and learning flourish alongside the applications of AI.””
Twitter · 09/21/23 · #425
Reach Capital · 09/11/23 · #424
EdTechInsiders · 09/13/23 · #424
ASCD · 09/01/23 · #423
“Draftback does a few helpful things for Google Docs. Most importantly, it turns a document's revision history—which we normally have to click through piecemeal—into a video that can be played at controllable speeds. When we watch it, we can actually see a student's every keystroke on a document. Draftback also provides a detailed breakdown of the document's history that includes the overall number of revisions, the number of "distinct writing sessions" (which Draftback defines as moments when there wasn't more than a 10-minute gap between revisions), the time and duration of each of those writing sessions, and which user made them!”
Hechinger Report · 09/04/23 · #423
One Useful Thing · 09/05/23 · #423
“Forget synonyms of words, find synonyms for paragraphs. I asked GPT-4 to Give me 20 vastly different variations on this [the paragraph above]. Make them as different as possible in style. label each style. Some of the more exciting examples of what it came up with:”
Academic DJ · 09/08/23 · #423
Learning On Purpose · 09/08/23 · #423
Education Next · 08/08/23 · #422
“It is wise to be skeptical of new technologies that claim to revolutionize learning. In the past, prognosticators have promised that television, the computer, and the Internet, in turn, would transform education. Unfortunately, the heralded revolutions fell short of expectations. There are some early signs, though, that this technological wave might be different in the benefits it brings to students, teachers, and parents. Previous technologies democratized access to content and resources, but AI is democratizing a kind of machine intelligence that can be used to perform a myriad of tasks.”
Harvard Graduate School of Education · 08/24/23 · #422
“Create a curriculum that is process-oriented, not product-oriented. AI should not be used to replace the “thinking” or ability to recall facts that product-oriented learning often asks of students. Rather, Dede and Cao suggest developing process-oriented learning approaches that encourage students to find answers with their own logic and reasoning. “We need to equip students for a world full of uncertainties and challenges and prepare them to do the complex work of problem-solving,” they write. “Embracing a messy process rather than becoming good at following “cognitive recipes.””
Colorado State · 08/31/23 · #422
“We hope this collection offers something for teachers with all levels of comfort with technologies—from teachers seasoned with digital writing technologies to teachers approaching the entire domain with trepidation. To that end, we have made the teaching resources in this collection as accessible as possible.”
OpenAI · 08/31/23 · #422
“We’re sharing a few stories of how educators are using ChatGPT to accelerate student learning and some prompts to help educators get started with the tool. In addition to the examples below, our new FAQ contains additional resources from leading education organizations on how to teach with and about AI, examples of new AI-powered education tools, and answers to frequently asked questions from educators about things like how ChatGPT works, its limitations, the efficacy of AI detectors, and bias.”
A.I. Educator · 09/02/23 · #422
EdWeek · 08/21/23 · #421
“Nearly 4 in 10 teachers expect to use AI in their classrooms by the end of the 2023-24 school year. Less than half as many say they are prepared to use the tools.”
A.J. Juliani · 08/24/23 · #421
“It has been nine minutes on this video, I now have ideas for an activity, a lesson plan connected to my standards, and a presentation, all setting students up.”
New York Times · 08/24/23 · #421
“First, I encourage educators — especially those who teach in high schools and colleges — to assume that 100 percent of their students are using ChatGPT and other generative A.I. tools on every assignment, in every subject, unless they’re being physically supervised inside a school building… My third piece of advice — and the one that may get me the most angry emails from teachers — is that teachers should focus less on warning students about the shortcomings of generative A.I. than figuring out what the technology does well.”
New York Times · 08/24/23 · #421
“As the school year begins, their thinking has evolved.”
Learning On Purpose · 08/25/23 · #421
“If you are already trying AI, show students what you’re doing. Ask them what they think. Tell them what you think. Your willingness to show your work models vulnerability and offers implicit permission to talk openly about using AI for learning… If you show your work, students are more likely to show theirs.”
AI for Education · 07/31/23 · #420
AI for Education · 08/01/23 · #420
LinkedIn · 08/01/23 · #420
AI for Education · 08/02/23 · #420
EdArXiv · 08/09/23 · #420
“To discuss these issues we think it is helpful to begin with a short explanation of how the Large Language Models… work; what they are capable of; how they are similar/different to human brains; and what the implications might be for human learning and motivations to learn. This is the focus of Part One of the paper. In Part Two, we explore four different scenarios for humanity and in particular, what each of these scenarios might mean for the future of learning and education. Finally, in Part Three, we present 13 recommendations that we think will help to ensure that AI becomes our greatest success, rather than a tangled mess.”
code.org · 08/15/23 · #420
The Academic DJ · 08/18/23 · #420
Berkeley · 06/13/23 · #419
Learning On Purpose · 08/11/23 · #419
“For me, the best way to make sense of AI is to use it. These tools are powerful, but they are imperfect. They have enormous potentials and pitfalls for our world, but the tools available to most of us are accessible and easy to use. They are already doing amazingly sophisticated work in fields like medicine and technology and business, but they also have simple and practical applications in our day-to-day lives. Open up a chatbot like ChatGPT or Bing or Bard or Claude and give a few of the below ideas a try. All of these tools are free to use (sometimes with limits).”
Politico · 05/01/23 · #417
“Encode Justice emerged from that ballot campaign with a focus on the use of AI algorithms in surveillance and the criminal justice system. It currently boasts a membership of 600 high school and college students across 30 countries. Revanur said the group’s primary source of funding currently comes from the Omidyar Network, a self-described “social change venture” led by left-leaning eBay founder Pierre Omidyar.”
EdWeek · 07/14/23 · #416
“Most telling is how few educators say they have received any professional development on how to incorporate AI into their work in K-12 education: Eighty-seven percent said they had received no such PD at the time of the survey.”
Arxiv · 07/21/23 · #416
“Detection tools for AI-generated text do fail, they are neither accurate nor reliable (all scored below 80% of accuracy and only 5 over 70%). In general, they have been found to diagnose human-written documents as AI-generated (false positives) and often diagnose AI-generated texts as human-written (false negatives).”
New York Times · 07/22/23 · #416
“Researchers have tried for more than half a century to design computer programs to make medical diagnoses, but nothing has really succeeded. Physicians say that GPT-4 is different. “It will create something that is remarkably similar to an illness script,” Dr. Rodman said. In that way, he added, “it is fundamentally different than a search engine.””
AI In Academia · 07/07/23 · #415
Cornell · 04/18/23 · #414
Slate · 06/26/23 · #414
“This summer, K–12 schools must get to work drafting academic policies governing the use of A.I. and facilitating professional development for teachers about the new technology. Educators need to address this trend head-on with students while simultaneously redesigning instruction and assessment in an age of A.I. We cannot start this work quickly enough.”
District Administration · 05/10/23 · #413
Chronicle of Higher Education · 05/12/23 · #413
Stanford · 03/09/23 · #411
Inside Higher Ed · 03/13/23 · #411
One Useful Thing · 02/01/23 · #410
Businesswire · 03/07/23 · #410
SmartBrief · 02/24/23 · #409
Twitter · 02/24/23 · #409
“Bing gave ChatGPT 70%. ChatGPT gave Bing an A”
Krupp Library · 04/12/18 · #408
Stanford Daily · 01/22/23 · #405
Twitter · 01/16/23 · #404
Mr. Pultz · 01/18/23 · #404
e-Literate · 12/16/22 · #403
Silicon Angle · 12/16/22 · #403
New York Times · 12/27/22 · #403
Hugging Face · 01/01/23 · #403
3 Quarks Daily · 01/02/23 · #403
NPR · 01/09/23 · #403
“To determine whether an excerpt is written by a bot, GPTZero uses two indicators: "perplexity" and "burstiness." Perplexity measures the complexity of text; if GPTZero is perplexed by the text, then it has a high complexity and it's more likely to be human-written. However, if the text is more familiar to the bot — because it's been trained on such data — then it will have low complexity and therefore is more likely to be AI-generated. Separately, burstiness compares the variations of sentences. Humans tend to write with greater burstiness, for example, with some longer or complex sentences alongside shorter ones. AI sentences tend to be more uniform.”