In a recent study conducted by researchers at Columbia University, a prototype AI writing-support tool named Scraft has been developed. This tool is designed to aid writing education by using recursive feedback mechanisms to encourage critical thinking.
Scraft is not just a simple text-generating AI; it’s a sophisticated tool that asks Socratic questions to users and provides personalized feedback throughout the writing process. This approach is designed to stimulate critical thinking and improve writing skills by engaging the writer in a recursive process of reflection and revision.
The researchers conducted a preliminary study with 15 students to evaluate the effectiveness of Scraft. The results indicated that the recursive feedback provided by Scraft was helpful in improving the students’ writing skills. However, the participants also noted that the feedback was sometimes factually incorrect and lacked context. This highlights the challenges of developing AI tools that can provide accurate and contextually appropriate feedback.
The researchers argue that AI writing-support tools should focus on preserving the recursive and thought-provoking nature of writing. This means that the AI should not just correct grammar and spelling errors, but also engage the writer in a dialogue that encourages reflection and revision.
Scraft could be particularly beneficial for multilingual learners. It can provide immediate, personalized feedback, which can be especially helpful for those who are learning English as a second language and may not have access to a human tutor. The Socratic questioning approach used by Scraft can also help multilingual learners to think critically in English, which is an important skill for academic writing.
However, it’s important to note that Scraft is still a prototype and further research is needed to improve its accuracy and contextual understanding. Despite these challenges, the development of Scraft represents an exciting step forward in the use of AI in education.
Thanks for reading. This site and all the work shared here are completely reader-supported. The best way to support it is to check out my recommendations or subscribe to my weekly newsletter.
The world of education, specifically K-12, is on the brink of a significant transformation. The catalyst? Generative AI tools. These tools, such as Large Language Models (LLMs) and ChatGPT, are heralding a new era of automation, promising to reshape how we approach administrative and teaching tasks in schools.
Generative AI tools are a generational leap in what we can automate with software. They are not just about replacing human effort but also about creating entirely new kinds of automation. The potential impact on jobs and people is profound, and the pace of change is rapid. For instance, ChatGPT has already amassed over 100 million users in just six months.
The world of education is no stranger to automation. Over the past two centuries, we’ve seen waves of automation that have eliminated certain jobs while creating new ones. This process, while sometimes disruptive, has ultimately led to increased prosperity and efficiency.
For school administrators and teachers, generative AI tools could automate many tasks, freeing up time for more strategic and student-focused activities. For example, these tools could automate administrative tasks such as scheduling, record-keeping, and communication with parents. They could also assist teachers with tasks such as grading, lesson planning, and even providing personalized learning support for students.
However, the adoption of these tools is not without challenges. The tools that people use to do their jobs are complicated and very specialized, embodying a lot of work and institutional knowledge. Replacing or automating any of these tools and tasks is not trivial. There’s a huge difference between an amazing demo of a transformative technology and something that a big complicated organization can use.
Moreover, while generative AI tools can answer ‘anything’, the answer might be wrong. They are not databases but pattern matchers. They can produce answers that fit the pattern of the question but may not be factually correct. This means that while they can automate many tasks, their outputs still need to be checked.
Despite these challenges, the potential benefits of generative AI tools in K-12 education are immense. They could lead to more efficient administration, more personalized learning, and ultimately, better educational outcomes for students. However, it’s important to remember that these tools are not a magic bullet. They are just another wave of automation, and their successful implementation will require careful planning, training, and adjustment.
In conclusion, generative AI tools hold great promise for automating tasks in K-12 education. However, their adoption will require careful planning and a clear understanding of their capabilities and limitations. As with any new technology, the key to success will be in how well we integrate these tools into our existing systems and processes, and how well we adapt to the new ways of working they enable.
FAQ
What is generative AI? Generative AI, including Large Language Models (LLMs) and ChatGPT, represents a significant change in what we can automate with software. It’s not just about replacing human effort but also about creating entirely new kinds of automation.
How fast is the adoption of generative AI tools like ChatGPT? The adoption is happening very rapidly. For instance, ChatGPT has amassed over 100 million users in just six months.
What is the potential impact of generative AI on jobs? Generative AI tools have the potential to automate many tasks, which could lead to job displacement. However, similar to previous waves of automation, they could also create new types of jobs.
What challenges are associated with the adoption of generative AI tools? The tools people use to do their jobs are complicated and very specialized, embodying much work and institutional knowledge. Replacing or automating any of these tools and tasks is not trivial. Additionally, while generative AI tools can answer ‘anything,’ the answer might be wrong as they are not databases but pattern matchers.
What is the potential of generative AI tools in the education sector? In the education sector, generative AI tools could automate many administrative tasks and assist teachers with tasks such as grading, lesson planning, and even providing personalized learning support for students.
What is the future of generative AI tools? The future of generative AI tools is likely to involve more automation, but also more integration with existing systems and processes. Their successful implementation will require careful planning, training, and adjustment.
What is the ‘Lump of Labour’ fallacy? The ‘Lump of Labour’ fallacy is the misconception that there is a fixed amount of work to be done and that if a machine takes some work, there will be less work for people. However, if it becomes cheaper to use a machine to make, say, a pair of shoes, then the shoes are cheaper, more people can buy shoes, and they have more money to spend on other things besides, and we discover new things we need or want, and new jobs.
What is the Jevons Paradox? The Jevons Paradox suggests that as technological progress increases the efficiency with which a resource is used, the total consumption of that resource may increase rather than decrease. This paradox has been applied to white-collar work for 150 years.
What is AGI (Artificial General Intelligence)? AGI refers to a type of artificial intelligence that is as capable as a human at any intellectual task. If we had AGI, it could potentially change everything, including overriding all the complexity of real people, real companies, and the real economy. However, as of now, we do not have AGI, and without that, we have only another wave of automation.
How can generative AI tools help in personalized learning? Generative AI tools can provide personalized learning support for students by adapting to each student’s learning style and pace. They can provide additional explanations, practice problems, and feedback, making learning more effective and engaging.
Can generative AI tools replace teachers? While generative AI tools can assist with tasks such as grading and lesson planning, they are not a replacement for teachers. Teachers play a crucial role in motivating students, managing the classroom, and providing emotional support, among other things. These are aspects that cannot be automated.
What is the role of generative AI tools in administrative tasks? Generative AI tools can automate administrative tasks such as scheduling, record-keeping, and communication with parents. This can free up time for school administrators to focus on more strategic tasks.
What is the difference between a database and a pattern matcher in the context of generative AI tools? While databases store and retrieve factual information, pattern matchers, like generative AI tools, generate responses based on patterns they’ve learned from data. This means they can produce answers that fit the pattern of the question but may not be factually correct.
What is the importance of careful planning and training in adopting generative AI tools? The successful implementation of generative AI tools requires careful planning and training. This is because these tools must be integrated into existing systems and processes, and users need to understand their capabilities and limitations.
What does it mean that generative AI tools are not a magic bullet? This means that while generative AI tools hold great promise, they are not a solution to all problems. Their successful implementation will require careful planning, training, and adjustment. They are just another wave of automation, and their impact will depend on how well we adapt to the new ways of working they enable.
What is the potential impact of generative AI tools on educational outcomes? By automating administrative tasks and assisting with teaching tasks, generative AI tools could lead to more efficient administration, more personalized learning, and, ultimately, better educational outcomes for students.
Thanks for reading. This site and all the work shared here are completely reader-supported. The best way to support it is to check out my recommendations or subscribe to my weekly newsletter.
The Innovating Pedagogy 2023 report, published by The Open University, explores ten innovations that have the potential to provoke major shifts in educational practice. The report is designed to guide teachers, policymakers, and educational technologists in making informed decisions about new forms of teaching, learning, and assessment.
Learning through Open Data: Open data is publicly available information that can be freely used, modified, and shared. The report suggests that open data can be used as a teaching tool to develop students’ data literacy skills, critical thinking, and understanding of complex issues.
Student-led Analytics: This innovation involves students in the process of collecting, analyzing, and using their own educational data to support their learning. It empowers students to take control of their learning and make informed decisions.
AI Teaching Assistants: AI teaching assistants can provide personalized learning experiences, answer students’ questions, and give feedback on assignments. They can support teachers by taking over routine tasks, allowing teachers to focus on more complex aspects of teaching.
Micro-credentials: Micro-credentials are digital certificates that recognize small amounts of learning or skills. They offer flexible pathways for lifelong learning and can be stacked to form a larger qualification.
Learning through Multisensory Experiences: This approach uses technologies such as virtual and augmented reality to provide immersive learning experiences. It can help students understand complex concepts and develop skills in a safe and controlled environment.
Humanistic Knowledge-Building Communities: These are online communities where learners and teachers collaboratively create knowledge. They foster a sense of belonging and support the development of critical thinking and problem-solving skills.
Learning from Robots: Robots can be used in education to support learning in various ways, such as teaching coding or providing social and emotional support to students.
Blockchain for Learning: Blockchain technology can be used to create secure, transparent, and tamper-proof educational records. It can also support the recognition of micro-credentials and facilitate the sharing of learning records across institutions.
Decolonizing Learning: This involves challenging the dominant Eurocentric perspective in education and incorporating diverse knowledge, cultures, and ways of knowing into the curriculum.
Action-Oriented Learning: This approach involves students in real-world problem-solving and social action. It develops skills such as collaboration, critical thinking, and civic engagement.
The report (available here) emphasizes that these innovations are not standalone solutions but should be integrated into a broader pedagogical strategy. It also highlights the importance of considering ethical issues, such as data privacy and the risk of AI bias, when implementing these innovations.
Thanks for reading. This site and all the work shared here are completely reader-supported. The best way to support it is to check out my recommendations or subscribe to my weekly newsletter.
Professor and friend John Nash co-hosts a podcast on all things online learning. In a recent episode, he shared his work on coaching ChatGPT to write more “human” and the results are… interesting…
While generative AI tools are very cool right now, they are a long way from being truly disruptive and overtaking the world.
Here’s what’s interesting. Scaffolding the prompts, defining perplexity and burstiness, and then prompting an explicit increase of those measures made the text “human” to GPTZero. Still, it also made the text ridiculously flowery and inflated. Kind of like when a master’s student thinks they are supposed to “sound academic.” It was so bad that the ChatGPT output was immediately suspect to my human eyes, even though GPTZero said it was likely written entirely by a human.
One of the teachers in my teacher fellowship asked me about introducing the process of developing and marketing a mobile app to her third-grade students. They are working on a unit about economics.
I’m a fan of the 5E lesson planning framework. If you’re unfamiliar, The 5E lesson plan is a widely used instruction model consisting of five phases: Engage, Explore, Explain, Elaborate, and Evaluate.
During the engage phase, teachers can use various strategies to encourage students to discover more about the topic, such as an attention-grabbing hook, a thought-provoking question, or a hands-on activity. The engage phase’s goal is to create an atmosphere of curiosity and excitement, as well as to set the tone for the rest of the lesson.
I thought about creating something for the first part of this 5E framework to get the students interested in the topic of mobile app development. Perhaps a little video of a mobile app developer giving students an overview of the process?
I know access to mobile app developers is easy to come by for schools in large cities, but my schools aren’t in a large city. And I don’t have any contacts immediately available to speak to a class, whether in-person or remotely.
Which, if you weren’t aware, is the situation that so very many schools deal with every day. Access to experts in some fields isn’t readily available.
But there’s a solution.
With a little help from a few AI tools, here’s what I did…
Writing a Video Script with AI
I began by using a tool I had just learned about called TeacherBot. The folks at TeacherBot aim to “build the tool every teacher deserves.” You enter an idea into the prompt space, and TeacherBot gives you a useful output.
The TeacherBot user interface
I aimed to create a script for an overview video of the mobile app development process. Here’s the prompt I gave TeacherBot to begin:
You are a software programmer. You need to introduce the concepts of coding an app for Apple’s iOS to a group of third-grade students in the United States. Create a story talking about the different phases of mobile app design and development.
I asked for a “story” and specified third-grade students to hopefully keep the response at an appropriate grade level for the students
Here’s what TeacherBot gave me in return:
TeacherBot response to prompt
Is it rudimentary? Yes. But, it is designed for a 3rd-grade audience. And it’s just an overview. If I had the learning objectives for the lesson or the unit, I might have been able to craft a better prompt or tweaked the response myself. I wanted a female speaker, so I should have indicated that in the prompt. However, that was an easy fix.
Now, I needed to create my speaker.
Creating a Speaker Image to Deliver the Script
To create an image for my speaker, I headed over to Midjourney and entered the following prompt:
minority woman, photography, depth of field f2.8 3.5, 25mm lens, 8k, ultra-realistic — v 4 — q 2 — v 4 — s 750 — uplight
I won’t go into the crafting of prompts for Midjourney here. Linus Ekenstam inspired this particular prompt, and he has a wealth of knowledge on his Twitter feed on crafting prompts.
Here are the initial results from Midjourney:
All of these images would work perfectly for this task. I wanted female minority representation for this project. Why? Because students need to know that jobs in computer programming and software development aren’t just for white men.
I choose this image from the four at random:
Now, I have my script and my speaker. Time for the final step.
Creating a Video with AI
For the final step, I turned to D-ID. After creating an account, I opened the “Create Video” tool and added the image of my speaker, and pasted my script in the script box.
D-ID interface
I changed the language settings to give the speaker an appropriate accent but still speak in English. Here is the final result:
Now, we have a little something to use to engage our third-grade students.
Is it perfect? Of course not. Do we still have a long way to go in the world of AI? Yes.
Will we ever get to use this tool to its fullest potential if we block it in school because we’re worried about cheating?
No.
AI is a tool, just like anything else. Use it. Be aware of what it can be used for, both for “good” and for “bad” — and keep in mind that what defines those terms is highly subjective.
Thanks for reading. You can get more articles like this as part of my weekly newsletters. I send out two per week: Tuesdays are deeper dives into education topics for paid subscribers, and every Friday, I share “10 Things” I think are cool and worth your time for free.
Yes, there’s more to say about ChatGPT and how we can use it in education. I’m doing my best to develop productive uses of this technology to override all the Negative Ninnys out there in Luddite Land.
I won’t have an answer for everyone, but there’s a decent chance one of the ideas I share here could inspire you to use ChatGPT yourself.
Why Some Teachers May Fear ChatGPT and other AI Tools
First of all, let’s start here: Anytime a new technology is introduced into the world, people are apprehensive. This phenomenon is not exclusive to the education world.
New technologies can be intimidating and scary because they represent change and the unknown. People may be hesitant to adopt new technologies because they are unsure how they will work or fit into their existing way of life.
Additionally, new technologies can sometimes create uncertainty about the future and its impact on society. This fear of the unknown can lead to resistance to change and skepticism about new technologies.
For instance, bicycles were once thought to cause insanity and something called “bicycle face” that was a danger to a person’s health. Weird, I know, but totally true. Fearing new technology isn’t new.
Technology enthusiasts have to keep things in perspective when talking about new tools. Not everyone gets as excited as we do about disruptive tools, and not everyone immediately sees the potential. Patience, brethren. Change comes slowly.
Teachers might be uncertain of how to effectively incorporate ChatGPT into their classrooms, making them feel uneasy. To ensure successful implementation, teachers should become well acquainted with the tool beforehand and develop a strategy tailored specifically to it.
Another reason could be that they are concerned about the potential for the tool to be misused by students, for example, by using it to cheat on assignments or exams. Also, some teachers may be concerned about the impact of using a tool like ChatGPT on students’ critical thinking skills and ability to engage in deep learning and reasoning.
I should probably also mention that every input is being tracked, especially during this “research” phase of ChatGPT. OpenAI doesn’t hide this at all; they tell you that everything you type and every conversation generated is being tracked when you sign up. They even ask you for feedback when you get a generated response. Be aware of what is going on when using a tool like this.
Finally, there may be concerns about the ethical implications of using a tool like ChatGPT in the classroom, particularly with regard to issues such as privacy and the potential for bias in the tool’s responses. We’ve already seen the impact of widespread disinformation and how social media can influence opinions.
On the other hand, teaching with ChatGPT offers educators many opportunities. The technology can be used to help students outline or organize their papers, and at its best, can serve as a powerful, fun-to-use digital tutor that works like an improved version of Google.
There are several ways that teachers can effectively use tools like ChatGPT in the classroom. One way is to use the tool to provide students with additional support and guidance as they work on assignments or projects.
For example, a teacher could use ChatGPT to help students brainstorm ideas, conduct research, or organize their thoughts. Additionally, the tool could provide personalized feedback to students on their work, helping them identify areas for improvement and providing them with guidance on making progress.
Another way that teachers can use ChatGPT effectively in the classroom is by incorporating it into activities designed to promote critical thinking and problem-solving skills. For example, the tool could be used as part of a debate or discussion activity, where students could use it to generate and evaluate different arguments and perspectives on a given topic.
This can help promote deep learning and reasoning and encourage students to think more critically about the information and ideas they encounter.
Additionally, teachers can use ChatGPT to create engaging and interactive learning experiences for their students. For example, the tool could be used to create interactive quizzes or games where students can ask questions and receive answers in real time. This can help keep students engaged and motivated and provide a more dynamic and interactive learning experience.
Overall, the key to effectively using tools like ChatGPT in the classroom is to be creative and to find ways to integrate the tool into teaching and learning activities in a meaningful and engaging way. This can help to enhance students’ learning experiences and promote critical thinking and problem-solving skills.
Thanks for reading. Get access to exclusive content and expert insights on technology, teaching, and leadership by subscribing to my newsletter. Stay up-to-date on the latest trends and join our community of professionals and educators worldwide.
The buzz around AI writing tools continues in the education world. Of course, there are several AI tools that we’re already using, whether in the classroom or not. We were using AI tools long before anyone thought about them invading our classrooms, but we didn’t think they had classroom applications.
But none have gotten the coverage that ChatGPT has gotten since its launch on 30 November 2022.
Um… I just had like a 20 minute conversation with ChatGPT about the history of modern physics. If I had this shit as a tutor during high school and college…. OMG.
I think we can basically re-invent the concept of education at scale. College as we know it will cease to exist.
1) If HS English dies, it’s not because of AI. It’s because standardized testing has shifted the focus of education from generating knowledge to regurgitating knowledge in a standard format. https://t.co/y0Ud9ffauS
Teachers are concerned that students will use @OpenAI's chatbot ChatGPT for plagiarism. But, given a national shortage of teachers, the tool could also be used as a co-teacher. I spoke to five teachers across the U.S. on how ChatGPT could impact education. https://t.co/rBdVgE93JH
I fear that the first response many educators will have is, “we have to block it right now.”
I understand teachers’ very valid concerns about any new technology tool, but blocking is horribly inefficient and the equivalent of burying our heads in the sand.
As tools proliferate, they become more and more difficult to block School IT departments get enough of these types of requests already, and in most cases, blocking one site only leads to students finding ten more that offer them the same access.
It’s not that I don’t think we need to have good conversations about the responsible usage of tools like ChatGPT. Without rails to guide the path, there is a strong possibility of misuse or poor usage. If there was ever a time when we needed more focus on digital citizenship and media literacy, I can’t think of one.
But we can talk about responsible usage of any tool in the classroom. The concept isn’t new. Before we had Google Docs, kids passed notes in class. The pen was once accused of the oncoming downfall of the education system.
How many times have you had to prevent your classroom from being invaded by ruler helicopters? Abusing tools in the classroom or, perhaps more correctly, using tools to avoid boredom in the classroom is nothing new.
So what do we do with new tools that are certain to disrupt the status quo?
My hope is that more of us have this outlook on new tools available to use in schools:
I wonder what Kenyans will do once they discover the magic of ChatGPT.
Some will abuse it. Others will use it to learn and grow. Only time will tell which side you pick.
Obviously, our classroom activities should challenge students to do more than regurgitate information. We should challenge students to create from their imagination.
We must strive for deeper learning in every classroom in every school.
If teachers design student-centered learning experiences that allow students to write with support in class, ChatGPT won’t be nearly as disruptive as some articles claim.
We should provide opportunities that stimulate their brain and make neural pathways come alive with dancing dreams of great design.
When we don’t embrace new technologies, we deny students options. We prevent them from learning about how their world is changing.
I love me some disruptive technology. There’s no point in beating our chest about how technology x has made y obsolete. The business world can not ignore disruptive technology or they will go out of business. As educators we are in the business of preparing students for THEIR future. The future for students includes AI (Artificial Intelligence).
But not only do we prevent students from experiencing new tools that can be very useful in their lives, but we also overlook what we, as teachers, can use these tools for to make our lives easier.
This #ChatGPT is BONKERS. I copied and pasted last week's parent email, copied upcoming events from our school calendar, and copied the standards we're teaching this week. I asked it to "write this week's parent email." It did it in 10 seconds. pic.twitter.com/QcTOvq2Chl
The emergence of AI education disruptors like ChatGPT reveal the need for more diverse teaching models. The COVID-19 pandemic was a catalyst, spurring teachers and administrators into action. We can’t return to “normal school” any more than we can ignore new educational advancements.
We must embrace change. We can’t move forward without it.
How Disruptive Will ChatGPT Be?
I introduced Minecraft: Education Edition to my school district last school year and made the statement in a school board presentation that it was likely the most disruptive tool I’d brought to the district.
But ChatGPT? Oh my. I hope it breaks more barriers and causes more people to rethink daily what they do in classrooms. We already know (or we should know) that students will use AI tools to write papers. I hope educators use it, and many other technologies, to completely redesign education for the future.
Thanks for reading. Get access to exclusive content and expert insights on technology, teaching, and leadership by subscribing to my newsletter. Stay up-to-date on the latest trends and join our community of professionals and educators worldwide.
You’re just learning about GPT-3, but folks are already working on GPT-4. Here’s what it might look like (emphasis on might)
The end-of-year recommended book lists are beginning to appear. I’ll have mine out closer to the end of December. Here are a few from reputable sources:
When AI is smart enough to write an essay, then what happens?
GPT3 is back in the news, because, as expected, it’s getting better and better. Using a simple chat interface, you can easily ask it a wide range of questions (write a 1,000 word essay about Clara Barton) that certainly feels like a diligent high school student wrote it.
Of course, this changes things, just as the camera, the typewriter and the internet changed things.
It means that creating huge amounts of mediocre material is easier than ever before. You can write a bad Seinfeld script in about six minutes.
It means that assigning rudimentary essays in school or average copywriting at work is now a waste of time.
But mostly it reminds us that attention and trust don’t scale.
If your work isn’t more useful or insightful or urgent than GPT can create in 12 seconds, don’t interrupt people with it.
Technology begins by making old work easier, but then it requires that new work be better.
I think it’s always important to consider the work we ask students to do in our schools. As my teacher cohort works through implementing the 4 Shifts protocol, we ask questions around deeper learning and authentic work like:
Is student work deeply rooted in discipline-specific and -relevant knowledge, skills, and dispositions?
Do learning activities and assessments allow students to engage in deep critical thinking and analysis?
Do students have the opportunity to design, create, make, or otherwise add value that is unique to them?
Is student work authentic and reflective of that done by experts outside of school?
Are students utilizing authentic, discipline-specific practices and processes?
Are students creating real-world products or performances for authentic audiences?
Of course, not every lesson or activity can be (nor should it be) an exercise in critical thinking and authentic, real-world application. But if our biggest concern about AI is whether or not students will use it to cheat, perhaps we have work to do on our classroom plans.
