APPLE

.: The AI upgrade cycle is here​

Summary ➜ Apple recently announced impressive new AI features called Apple Intelligence at WWDC 2024 that will be able to take actions based on user data, manage notifications, and rewrite text. However, these features will only be available on the latest, most expensive iPhone 15 Pro and Pro Max models, as well as iPads and Macs with M1 chips or newer. Powerful new AI features are becoming the latest way for tech companies to encourage faster device upgrade cycles, though the true hardware limitations aren’t always transparent. This may compel even happy users of older devices to upgrade sooner than planned.

Why this matters for education ➜ So, Apple makes a considered move with AI. The modest features announced such as only being able to rewrite text, instead of endlessly generating new copy, are grounded in practicality and ensure AI is useful, not just potentially useful. There are two stand out reasons why the Apple announcement from WWDC 2024 matters for education:

1. Privacy and security – as I called out a while ago Apple is putting private, secure and practical AI use in the forefront. They are doing lots of interesting stuff with on-device computation and private cloud integration. These standards will ripple out across the tech ecosystem and will impact how we, and our students, work with AI tools.
2. Making it easy to try AI – combine the above with the potential market for on-device AI something in the region of a billion devices and you see the impact it might have. There is an AI bubble and not as many people have tried AI tools as you might think – Apple’s market position will make it easy for iOS users to try new AI capabilities.

When we zoom out and see how these changes will ripple across the system of expectations and users, it is clear this matters for all AI users.

.: Other News In Brief

​

🚫 Microsoft’s all-knowing Recall AI feature is being delayed

👀 Former head of NSA joins OpenAI board

⚡️ Tesla investors sue Elon Musk for launching a rival AI company

💰 Amazon says it’ll spend $230 million on generative AI startups

👁️‍🗨️ GPTZero’s founders, still in their 20s, have a profitable AI detection startup, millions in the bank and a new $10M Series A

🍎 Apple joins the race to find an AI icon that makes sense

🎬 AI is the sixth great revolution in filmmaking (and maybe the most important)

🇪🇺 Meta says European regulators are ruining its AI bot

⛑️ Google Gemini proves a better health coach than humans

🍿 ‘We don’t need Sora anymore’: Luma’s new AI video generator Dream Machine slammed with traffic after debut

​

What’s on my mind?

.: Did ChatGPT break open the Overton Window?

​

The Overton Window shows a range of ideas that people think are okay to talk about and consider. Ideas inside the window are seen as normal and acceptable, while ideas outside the window are seen as too extreme or not acceptable to most people. Named after Joseph Overton, the window is commonly used to illustrate how public perception of what is politically feasible can shift over time. However, it can also be used to consider technology adoption.

The Overton Window can move as people’s opinions change. Ideas that were once seen as too extreme can become more acceptable and move into the window. For example, think about how people’s views on things like school uniforms, cell phone use in school, or environmental protection have changed over time. Famous people, leaders, and events can change the Overton Window. When influential people talk about new ideas, they can help those ideas become more acceptable.

By Hydrargyrum CC BY-SA 2.0

Understanding the Overton Window helps us see why certain ideas become popular and others don’t. It shows us how public opinion changes and why some topics are discussed more than others. By knowing about the Overton Window, you can better understand how society changes over time and how new ideas become part of our everyday lives.

One such event was the public release of ChatGPT by OpenAI—according to some recent reports, the OpenAI board did not know about the release! I have been wondering about the lasting impact on public acceptance and perception of the release, and the ensuing wave of other large technology company announcements. How much traction has this really gained with the majority of the public? Are people ready to accept and integrate AI tools like ChatGPT into their everyday routines, or do they still see them as too advanced or risky?

Apple’s modest feature release for new iPhones will also shift the window of acceptance, due to the easy access to AI tools grounded in everyday practical use, as well as the vast market share Apple has with mobile devices. I can see on the near horizon the window shifting further as more people use AI tools on their mobile devices and in simple chat interfaces. Although right now there is a bubble of acceptance, use, and to some extent positive sentiment, the growing acceptance of AI use is not far off.

The ChatGPT release may have forced the window ajar, but it’s everyday use on mobile devices that might keep it open.

~ Tom

Prompts

.: Refine your promptcraft

​

During my time in India this week, working alongside some talented technology integration teachers, I was struck by how ChatGPT sparked a flurry of creativity for some, during a design challenge.

This experience reinforced my belief that AI can be an incredible ally for expanding our ideas – if we approach it right.

The key? Start with your own initial spark, then let AI push those ideas further, in directions you might not have imagined.

We all have biases and blind spots; AI can help us see beyond them – or at the least, begin to notice them.

Below, I’ve shared a sample prompt that takes this approach. Give it a whirl!”

:. .:

Remember to mak​e this your own, try different language models and evaluate the completions.

Learning

.: Boost your AI Literacy

​

INTRO

.: Explained: Generative AI​

The MIT News article provides a comprehensive introduction to generative AI, explaining its definition, mechanisms like Markov chains, GANs, diffusion models, and transformers, and diverse applications such as text and image generation. It also addresses challenges like bias, plagiarism, and worker displacement, offering insights into both the technology and its implications.

​

ASSESSMENT ​
​.: Do Students Want AI Writing Feedback?​

In a recent experiment, Mike Kentz tested AI-generated feedback on student essays alongside his own feedback. Students found AI feedback clear and useful but preferred personalised feedback from their teacher. The AI tool, while efficient, often provided generic suggestions and struggled with nuanced critique. The study highlights the balance needed between AI’s efficiency and the irreplaceable human touch in education, suggesting that grading should focus more on the writing process than just the final output.

​

COURSE ​
​.: TAFE NSW | Introduction to Artificial Intelligence​

A free beginner-friendly course from CSIRO and TAFE NSW in Australia, that covers real-world applications and terminology without needing prior programming knowledge. The course offers insights from industry experts and covers topics like machine learning and natural language processing in a 2.5-hour online, self-paced format.

Ethics

.: Provocations for Balance
​

AI Access and Inequality:

• Provocation: Apple’s decision to restrict advanced AI features to its latest and most expensive devices raises concerns about exacerbating the digital divide. This divide refers to the gap between those who have access to information and communication technologies and those who don’t. By limiting advanced AI tools to those who can afford high-end products, Apple could create a situation where wealthier individuals and institutions have a significant advantage in areas like education, healthcare, and job opportunities. This could widen the gap between the haves and have-nots in society, leading to further social and economic inequality.
• Dilemma: Should technology companies prioritise short-term profits by reserving cutting-edge AI features for expensive products, or should they take on a greater responsibility to ensure equitable access to these powerful tools for everyone, regardless of income level?

AI and Privacy:

• Provocation: Apple’s integration of on-device AI offers a potential double-edged sword. While it enhances privacy by processing data directly on user devices, it also deepens user lock-in within Apple’s ecosystem. This ecosystem refers to the interconnected products and services offered by a single company. By keeping data processing on-device, Apple reduces the need to send user information to the cloud, potentially making it more difficult for hackers or other unauthorized parties to access. However, this approach also means that users become more reliant on Apple’s hardware and software, potentially limiting their choices and making it more difficult to switch to other platforms in the future.
• Dilemma: Is the trade-off between enhanced privacy through on-device computation and the potential for increased data lock-in with a single company justified? How can we strike a balance between protecting user privacy and fostering a healthy level of competition within the tech industry?

Personalized Learning and Human Connection:

• Provocation: The rise of AI-powered personalised learning raises a critical question: can AI effectively replicate the importance of human connection in education? Proponents of AI-driven approaches argue that these systems can tailor learning experiences to individual student needs, potentially leading to improved outcomes. However, as Dan Meyer points out, AI chatbots lack the ability to foster the social relationships that are crucial for effective learning. These human relationships provide students with opportunities for collaboration, communication, and critical thinking, all of which are essential for developing well-rounded individuals.
• Dilemma: Should the education system prioritise the scalability and efficiency promised by AI-powered personalised learning, even if it comes at the expense of fostering essential human connections?

.: :.

Inspired by some of the topics this week.

Got feedback for me?

​

If you have any other kind, specific and helpful feedback, please reply to this email or contact me at tom@dialogiclearning.com​

​
.: Tom Barrett

US EDU

.: All U.S. K-12 educators to get free access to Khanmigo as Microsoft partners with Khan Academy​

Summary ➜ This collaboration also aims to use Microsoft’s Phi-3 models to improve AI-driven mathematics tutoring. The initiative is designed to offer personalised learning experiences and help teachers create more effective educational plans, thereby making teaching more sustainable and enjoyable. Additionally, Khan Academy will integrate more of its content into Microsoft Copilot and Teams for Education to expand access to educational resources.

Why this matters for education ➜ This announcement, from Microsoft’s developer conference, takes my top billing in this week’s issue because the (potential) direct impact on educators (at least in the US).

But after the fanfare, what questions do we have? I’ll go first:

• What do we know about the effectiveness of Khanmigo or any similar agent tutor tools? Has it improved?
• How will educators and students be able to control the experience?
• Is Khanmigo just the next version of Clippy?
• Where’s the space for other tools, and how can students get hands on to build their own?
• Is it me or is there something strange about putting “Teachers are super overworked” at the centre of the rationale for these tools? (more on this below)

Lots to think about here and it will be interesting to hear from Sal Khan in August at this year’s big tech and education conference in Australia.

.: Other News In Brief

​

🧠 Anthropic releases a research paper looking inside the black box of AI

👑 NVIDIA Shows Once Again Who is the Real King of Generative AI

🔍 EU’s ChatGPT taskforce offers first look at detangling the AI chatbot’s privacy compliance

🧑‍🤝‍🧑 Meta’s new AI council is composed entirely of white men

❌ Google Search’s “udm=14” trick lets you kill AI search for good

🎵 Spotify experiments with an AI DJ that speaks Spanish

🚀 NVIDIA says 20,000 GenAI startups are now building on its platform

💰 Elon Musk’s xAI raises $6 billion to fund its race against ChatGPT and all the rest

💵 Amazon is Considering $20 Monthly Subscription for GenAI Enhanced Alexa

📰 OpenAI partners with Wall Street Journal publisher News Corp.

​

What’s on my mind?

.: A Productivity Paradox

​

How is the great fanfare, hype, and clamour around AI tools in education a distraction from the working conditions in our education systems?

I may not be able to answer this question from every angle, but I have a hunch it’s the right query.

My unease stems from the recent surge in big tech announcements and the relentless marketing spin touting “productivity” gains for educators using these AI tools and systems.

Long-time readers will know that I have been highlighting the real workload challenges in our schools for some time. But something seems off-balance when Sal Khan, the founder of Khan Academy, says while introducing Khanmigo,

Is that the foundation stone from which we are meant to launch into a new era of super-productivity and super-creativity?! Super?!

This feels like a distraction from the rotten conditions that created the unreasonable work expectations in the first place.

We all need a lift, and technology like AI gives us a faster and more efficient way to complete our tasks. Don’t misinterpret my sentiment here—go for it and get stuck in! Explore these AI tools, try them out, and see how they work for you.

Take the draft lesson plans, get the report comments, and adapt the mountain of emails you need to send.

Just promise me once you have gleefully hacked and slashed your way through your to-do list and tamed your inbox, you will pause.

Pause and ask this simple question about your productivity:

~ Tom

Prompts

.: Refine your promptcraft

​

Another example of the recommended prompts shared by Anthropic, the research lab behind the Claude family of AI models.

I have been writing lesson plans for a major AI Literacy programme recently and it got me thinking about what it was like to learn how to write lesson plans as a young student teacher.

There is something powerful and clarifying in the process of planning a lesson. Its a complex orchestration of concepts and kids. Learning (or lesson) design is a special type of skill and mindset.

With all that said it is fascinating to see AI systems predict their way to an average lesson plans.

Here is the Lesson Planner prompt from Anthropic.

Edit the subject information to anything you want and adapt the task prompt too. Note the promotion of active learning and critical thinking. Here are some examples of how different AI systems complete the task.

Part of the response from OpenAI’s GPT-4o

Here we have a Think Pair Share activity as an engager and a transition into direct instruction from OpenAI’s GPT-4o.

Part of the response from Gemini via Workspace

Gemini needs to work on what a Hook is intended to do! I don’t think this is enough for Year 7s!

Worth noting here that these models have no understanding of the concept of designing a hook or engagement activity for a lesson. It just predicts the next most likely word.

Anthropic’s older Claude-2-100k model suggested a 3-2-1 exit ticket strategy for finishing the lesson. Most of the models seemed to like the exit ticket idea.

Part of the response from Claude-2-100k

My recommendation with any of these AI systems for learning design is

1. Take responsibility.
2. Refine the prompts to suit your context, community and class.
3. Amplify the pedagogical approach you want.
4. Generate LOADS of examples and push the tools to do weird things. Filter for ideas you can build around.
5. Break down the learning design process (Plan) into smaller chunks.
6. Build a bot which can do all of these things, so you don’t have to keep prompting.

:. .:

Remember to mak​e this your own, try different language models and evaluate the completions.

Learning

.: Boost your AI Literacy

​

INTRO

.: Explained: Generative AI​

The MIT News article provides a comprehensive introduction to generative AI, explaining its definition, mechanisms like Markov chains, GANs, diffusion models, and transformers, and diverse applications such as text and image generation. It also addresses challenges like bias, plagiarism, and worker displacement, offering insights into both the technology and its implications.

​

ASSESSMENT ​
​.: Do Students Want AI Writing Feedback?​

In a recent experiment, Mike Kentz tested AI-generated feedback on student essays alongside his own feedback. Students found AI feedback clear and useful but preferred personalised feedback from their teacher. The AI tool, while efficient, often provided generic suggestions and struggled with nuanced critique. The study highlights the balance needed between AI’s efficiency and the irreplaceable human touch in education, suggesting that grading should focus more on the writing process than just the final output.

​

COURSE ​
​.: TAFE NSW | Introduction to Artificial Intelligence​

A free beginner-friendly course from CSIRO and TAFE NSW in Australia, that covers real-world applications and terminology without needing prior programming knowledge. The course offers insights from industry experts and covers topics like machine learning and natural language processing in a 2.5-hour online, self-paced format.

Ethics

.: Provocations for Balance
​

Scenario 1: A New Hope

In a dystopian future, an entire generation has learned math solely from tools like Khanmigo. Though they test well, they are mathematical zombies, unable to innovate or problem-solve without an AI feeding them steps. Society stagnates as complex challenges like climate change and disease outbreaks fester, untackled by minds trained in rigid algorithmic thinking. Once hailed as a boon to education, AI systems are now seen as an intellectual prison. An underground resistance of human maths teachers arises, working in hidden analogue classrooms to cultivate the creative mathematical spark the world desperately needs. But they are hunted by the Algorithm Enforcement Agency, which seeks to stamp out any challenge to the AI teaching regime…

​

Scenario 2: Voice Theft

In a world where AI can clone voices with near-perfect accuracy, personal privacy becomes a relic of the past. Celebrities and ordinary citizens find their voices used in unauthorised ways—advertisements, political speeches, and even criminal activities. The boundaries of consent are blurred, as anyone’s voice can be synthesised and manipulated without their knowledge or approval. This leads to widespread paranoia and identity crises, as people can no longer trust what they hear. The lines between reality and AI-generated deception become indistinguishable, causing a breakdown in social trust and personal security.

​

Scenario 3: The Lonely Generation

Fast-forward a decade. Universities without professors are now the standard, with ‘mega-courses’ of thousands of students ‘taught’ by AI. Overworked faculty have been reduced to ‘AI wranglers,’ providing prompts to lifeless algorithms. The college experience has transformed into a bleak imitation, with students never interacting with a human instructor. But the most disheartening are those oblivious to the difference, the Zoomers who’ve never experienced learning as anything other than staring at a screen. The ‘Lonely Generation,’ raised by algorithms, grapple with forming human connections. A sombre new academic department emerges: ‘Crisis Counselling and AI Addiction Recovery’”

.: :.

Inspired by some of the topics this week and dialled up.

How would you rate this issue of Promptcraft?

(Click on your choice below)
​

​⭐️⭐️⭐️⭐️⭐️ Loved it!
​​
​⭐️⭐️⭐️ Good
​​
​⭐️ Poor​

​

If you have any other kind, specific and helpful feedback, please reply to this email or contact me at tom@dialogiclearning.com​

​
.: Tom Barrett

ACCESS

.: OpenAI releases GPT-4o, the most advanced model, for free​

Summary ➜ OpenAI has released GPT-4o, a new, free AI model that surpasses GPT-4 by a significant margin. This model is groundbreaking as it offers multimodal capabilities, including text, audio, video, and image processing. By making GPT-4o free, OpenAI has disrupted the AI landscape and set a new standard for competitors.

Why this matters for education ➜ I decided to curate this as the lead story because it concerns access to the most powerful AI capabilities.

There has always been a big difference between the GPT-3.5 model and GPT-4, the latter being part of a paid subscription. With this update, any educator or student with a free account can explore the most powerful model.

When using GPT-4o, ChatGPT Free users will also get access to features such as:

• Experience GPT-4 level intelligence
• Get responses from both the model and the web
• Analyse data and create charts
• Chat about photos you take
• Upload files for summarising, writing or analysing
• Discover and use GPTs and the GPT Store
• Build a more helpful experience with Memory

Of course we need safe connections to these powerful tools and we should question what the trade-off might be in this scenario. But it is exciting to think every teacher with a connected device across the globe can access this powerful and flexible tool.

Read more of my reflections below about the OpenAI announcement and the updates to Voice mode.

.: Other News In Brief

​

🤖 Google strikes back at OpenAI with “Project Astra” AI agent prototype

📁 ChatGPT now lets you import files directly from Google Drive, Microsoft OneDrive

🌍 Microsoft’s carbon emissions up nearly 30% thanks to AI

📸 Instagram co-founder joins Anthropic as chief product officer in fight against OpenAI

🎥 Google unveils Veo, a high-definition AI video generator that may rival Sora

🇺🇸 US Senators unveil proposed $32 billion roadmap for regulating AI

🎵 Sony Music warns tech companies over ‘unauthorised’ use of its content to train AI

💰 Hugging Face is sharing $10 million worth of compute to help beat the big AI companies

📱6 Things Announced at Google I/O, One That Mattered A Lot, and One Mystery

🖼️ Google still hasn’t fixed Gemini’s biased image generator

​

What’s on my mind?

.: Closer Synthetic Conversation

​

This week has been filled with exciting tutoring demonstrations and educational AI announcements from OpenAI and Google, with more likely to come from Microsoft and Apple soon. What’s been on my mind, though, is a particular improvement in the responsiveness of voice models demonstrated by OpenAI’s new Voice mode with GPT-4o and Google’s Gemini Live demo in Kings Cross (which also featured some intriguing glasses!).

​

While shaving milliseconds off response times might not seem like much, the demonstrations we’ve seen, especially from OpenAI, show that AI systems are integrating more fluidly into human dialogue patterns. This is a significant change. Reduced latency—the processing or wait time—means we are more engaged. We might stick around in voice chats longer and feel that these systems are more helpful.

There was even a deliberate demonstration of interrupting the AI system as it spoke, highlighting this improvement in user experience. Both OpenAI and Google showcased how their systems will utilise live vision through the camera, tapping into spatial data as a rich source of new information for training once research labs have exhausted other data sources.

If you haven’t experienced using the frontier AI systems in voice chats, I highly recommend giving it a try. It is the most fluid and natural way to interact, and it seems that the vision of these leading labs places voice control at the forefront.

Putting aside the technical advancements for a moment, let’s consider what this means for us as humans. The proximity to human dialogue has changed in these demonstrations. The speed of interaction is crucial, but you’ll also notice the intonation, the personality, and the informality in the communication style. Remember, all of this is designed, intentional, and deliberate.

I’m curious to know what it feels like to experience synthetic conversational fluency. It’s already quite fascinating, and I’m intrigued by how millisecond changes can profoundly affect our experience of synthetic digital relationships.

​

~ Tom

Prompts

.: Refine your promptcraft

​

This week I am featuring one of the recommended prompts shared by Anthropic, the research lab behind the Claude family of AI models.

Second-grade simplifier

Make complex text easy for young learners to understand.

You could use the following System prompt in your own GPT, Gem or Bot. Or use both within a normal chatbot session.

System Prompt

Here is some example text from the technical research paper linked below about LearnLM.

And here is the 3 paragraph result using our simplifier prompt with GPT-4o.

:. .:

Remember to mak​e this your own, try different language models and evaluate the completions.

Learning

.: Boost your AI Literacy

​

SPATIAL AI

.: With Spatial Intelligence, AI Will Understand the Real World | Fei-Fei Li | TED​

​

​

INTRO ​
​.: AI Explained: It’s maths, not magic (a simple guide) from Google and Australia’s CSIRO

​

​

RESEARCH ​
​.: Towards Responsible Development of Generative AI for Education: An Evaluation-Driven Approach​

Dig into the research and development of Google’s LearnLM model.

Ethics

.: Provocations for Balance
​

Scenario 1: The Perfect Listener

OpenAI releases a groundbreaking update to its AI, enabling unprecedented conversational fluency and emotional responsiveness. In classrooms, these AI assistants become the go-to for students, offering instant feedback, personalised tutoring, and empathetic listening. Teachers notice a significant increase in student engagement and academic performance. However, as students grow more attached to these AI companions, they begin to confide in them more than in their human teachers or peers. This deep reliance raises concerns about the erosion of human relationships and the potential for AI to misinterpret or misuse sensitive information. The school board must decide whether to continue using these highly responsive AI assistants, which clearly benefit education, or to limit their use to ensure that human connections and emotional intelligence remain central to students’ development.

​

Scenario 2: Human Touch

In the midst of a nationwide teacher shortage, a small town’s school district struggles to fill vacant teaching positions. To cope, the district relies heavily on the remaining human teachers, who are overwhelmed and burnt out. Desperate for a solution, the school board introduces advanced AI teaching assistants powered by GPT-4o to support the overworked staff. Initially, the AI assistants help alleviate the workload, allowing teachers to focus on more meaningful interactions with students. However, as the AI takes on more responsibilities, it begins to undermine the authority and expertise of the human teachers. The community must decide whether to continue relying on AI to fill the gap, risking the devaluation of the teaching profession, or to seek alternative solutions that might be less efficient but preserve the integrity and importance of human educators.

​

Scenario 3: Eye of the Beholder

A leading tech company releases an innovative AI system called “SpatialSense,” designed to enhance learning by utilising spatial data from students’ environments. This AI collects data through cameras and sensors, analyzing everything from classroom layouts to students’ body language and social interactions. It customizes lessons in real-time, tailoring content to each student’s unique spatial and social context. Initially, the results are impressive: engagement and comprehension rates soar. However, as the AI becomes more integrated into the educational system, concerns arise about privacy and the ethical use of such detailed personal data. When a student discovers that the AI has recorded sensitive moments without consent, the school system faces a dilemma.

.: :.

Inspired by some of the topics this week and dialled up.

How would you rate this issue of Promptcraft?

(Click on your choice below)
​

​⭐️⭐️⭐️⭐️⭐️ Loved it!
​​
​⭐️⭐️⭐️ Good
​​
​⭐️ Poor​

​

If you have any other kind, specific and helpful feedback, please reply to this email or contact me at tom@dialogiclearning.com​

​
.: Tom Barrett

HARDWARE

.: Apple teased AI improvements, including the M4’s neural engine, at its iPad event​

Summary ➜ Apple highlighted AI features, including the M4 neural engine, at its recent iPad event. The company showcased AI-powered tools like visual lookup and live text capture on the new iPad Air and Pro models. Apple hinted at future AI advancements for developers in iPadOS.

Why this matters for education ➜ Apple are yet to reveal their hand regarding AI strategy, and by all accounts we will hear more in their developer event in June. When you consider these device upgrades, chip improvement and the challenge of devices dedicated to AI – perhaps mobile phones and tablet technology will see a new wave of development from AI.

Using AI tools on-device, instead of via cloud based services, is likely to offer performance benefits, greater flexibility, as well as, improved standards in privacy and safety which is a key component for education implementation.

At the very least, I think we will see more personal control and new data privacy standards, which the AI ecosystem will have to engage with.

.: Other News In Brief

​

📸 OpenAI working on new AI image detection tools

🕵️‍♂️ Microsoft launches AI chatbot for spies

🔍 OpenAI to steer content authentication group C2PA

📚 Audible deploys AI-narrated audiobooks

🐋 Sperm whale ‘alphabet’ discovered, thanks to machine learning

🛡️ How VISA is using generative AI to battle account fraud attacks

🤖 Apple poaches AI experts from Google, creates secretive European AI lab

📲 Siri for iOS 18 to gain massive AI upgrade via Apple’s Ajax LLM

📱 Anthropic finally releases a Claude mobile app

💬 Google adds AI conversation practice for English language learners

​

What’s on my mind?

.: US-Centric Bias and its Impact

​

My recent collaboration with teachers from across Scandinavia – Norway, Denmark, Sweden, and Finland – reminded me of a critical concern within the growing use of AI in education. The issue? The potential for bias and cultural insensitivity within AI tools, particularly language models (LLMs).

Many leading AI companies and the datasets used to train their AI systems are rooted in the United States. This US-centric origin can create limitations – the AI may lack a nuanced understanding of cultural differences, leading to biases in its output. It highlights the need for a broader, more inclusive approach to AI development.

This issue reminds me of the “mirrors, windows, and doors” model often used in education. This concept emphasises the importance of the following for students:

• Mirrors: Seeing themselves reflected in the learning materials.
• Windows: Offering insights into different perspectives and cultures.
• Doors: Opening up opportunities for engagement with the world on a larger scale.

In the same way, the AI tools used in our classrooms should also embrace these principles.

A recent example of this bias can be seen in image generation tools. Meta AI, a widely used platform, came under fire for consistently depicting Indian men in turbans. (See above for the story)

While turbans are a significant part of Indian culture, their overwhelming presence in the AI’s output ignores the vast diversity of clothing and ethnicities within India. This highlights the need for AI developers to incorporate more geographically and culturally diverse datasets during training.

Educators have a vital role in driving change. We need to champion the development of more inclusive, culturally sensitive AI.

​

~ Tom

Prompts

.: Refine your promptcraft

​

During my current visit to Sweden, where I am working with teachers, I have found it fascinating to learn about the various ways they have been incorporating AI tools into their work.

One particular example that seems to strike a chord with educators across different countries is the use of AI tools to refine, adapt, and improve email communication with parents.

Although I never personally experienced the need to email parents during my teaching career, many teachers I collaborate with have expressed the pressure and anxiety they feel when communicating via email.

They often worry about striking the right tone, being clear and concise, and maintaining a professional yet approachable demeanour.

A helpful promptcraft technique to address this challenge is to develop a short style guide based on your own written content.

By analysing your previous emails and identifying the key elements of your communication style, you can create a set of guidelines that reflect your unique voice and approach.

Then, when crafting prompts for AI tools, you can incorporate these style guidelines to ensure that the generated content aligns with your personal communication style.

To give you an example, here’s a glimpse into my email writing style:

​

To create your own writing style guide just use a prompt similar to the example below:

By providing the AI tool with this style guide as part of the prompt, you can maintain consistency in your communication and reduce the time and effort required to compose emails.

:. .:

Remember to mak​e this your own, try different language models and evaluate the completions.

Learning

.: Boost your AI Literacy

​

GLOSSARY

.: The A-Z of AI: 30 terms you need to understand artificial intelligence​

​

HIGHER-ED ​
​.: University Students Concerned About AI Skills​

University-bound students are worried about how AI usage by others may affect their academic and career opportunities. A study of 1,300 students shows that many see AI as both helpful and concerning, with concerns about ethics and competitive disadvantages.

​

FACES ​
​.: Can you spot which of these faces is generated by AI? Probably not — here’s why​

Experts say it’s becoming harder to tell AI-generated faces from real ones. People often mistake AI faces as real due to advancements in technology. Media literacy and awareness are crucial to navigate this new landscape.

Ethics

.: Provocations for Balance
​

Scenario 1: The “All-American” Student

A school adopts an AI-powered “virtual tutor” advertised to provide personalised learning paths. Soon, students from immigrant families and international students report getting recommendations heavily biased towards Western history, US-centric examples, and subtly promoting American cultural norms and ideals over their native ones.

​

Scenario 2: The “Perfect” Uni Application

A new AI tool goes viral, promising to “optimise” university essays, suggesting not just edits but rewriting sentences to appeal to what it claims are admissions officers’ preferences. Counselors find that AI-driven revisions favour stories of overcoming hardship that conform to American narratives of “grit,” potentially erasing nuanced experiences of marginalised students.

​

Scenario 3: When Translation Goes Wrong

To better communicate with parents, a school adopts an AI-powered translation tool for emails and newsletters. Immigrant parents soon complain that translations are not just inaccurate, but convey disrespect or perpetuate stereotypes about their cultures. Turns out, the AI model wasn’t trained with nuanced understanding of cultural idioms.

Inspired by some of the topics this week and dialled up.

How would you rate this issue of Promptcraft?

(Click on your choice below)
​

​⭐️⭐️⭐️⭐️⭐️ Loved it!
​​
​⭐️⭐️⭐️ Good
​​
​⭐️ Poor​

​

If you have any other kind, specific and helpful feedback, please reply to this email or contact me at tom@dialogiclearning.com​

​
.: Tom Barrett

AI IMPACT

.: Texas is replacing thousands of human exam graders with AI​

Summary ➜ The Texas Education Agency (TEA) 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.

The new system uses natural language processing and after a December 2023 trial, will be applied to the grading of open ended questions on the State of Texas Assessments of Academic Readiness (STAAR) exams. These are the standardised assessments for all public school students in grades 3–8 and specific high school courses.

Why this matters for education ➜ A story which seems to have so much of the relevant topics woven through it. We’ve got workforce impact, propping up a standardised testing regime, the potential for bias in AI grading, and the big question of whether AI can truly understand and evaluate student responses in a nuanced way that reflects deep learning, rather than just surface-level answer matching.

Relying on AI to assess student learning on high-stakes standardised tests is already raising concerns. How can we ensure the AI system grades fairly and accurately across different student demographics? What recourse will students and teachers have to appeal AI-awarded scores they believe are wrong? There’s a risk that AI grading entrenches or exacerbates existing biases and inequities in standardised testing.

.: Other News In Brief

​

🇺🇸 A new US bill wants to reveal what’s really inside AI training data

🤖 Mentee Robotics unveils an AI-first humanoid robot

🦁 Brave launches a real-time privacy-focused AI answer engine

⚡ Power-hungry AI is putting the hurt on global electricity supply

🖱️ Logitech wants you to press its new AI button​

🎓 Stanford report: AI surpasses humans on several fronts, but costs are soaring

🇬🇧 UK mulling potential AI regulation​

🎥 Adobe to add AI video generators Sora, Runway, Pika to Premiere Pro

🎯 X.ai Announces Grok-1.5V Multimodal Foundation Model and a New Benchmark

🌐 Google’s new technique gives LLMs infinite context

​

What’s on my mind?

.: Have we made progress?

​

The AI hype train keeps rolling, but are we getting anywhere? As an educator, I am increasingly frustrated with the repetitive discussions in edtech networks and the constant influx of marginally better AI systems and policy updates.

But what real progress have we made?

Let’s look at one characteristic of AI systems and whether much has shifted for us in education over the last few years.

AI systems are opaque.

AI opacity refers to the lack of transparency surrounding what happens between submitting a prompt and receiving a response and the training data used by AI companies.

This “black box” nature of most commercial AI systems is a significant concern within the educational context, as it directly impacts our ability to trust and effectively integrate these tools into our teaching and learning processes.

There are a plethora of free resources and courses available to increase our understanding. Jump ahead to the AI Literacy section below for a great example.

Recent controversies, such as OpenAI allegedly scraping and using YouTube content against their terms of service and Google engaging in similar practices for their own AI training, highlight the ongoing lack of transparency.

Kevin Roose and Casey Newton explore this topic in the latest edition of Hard Fork.

​

Looking back at my in-depth exploration of AI attribution and provenance last June, it’s striking how little has changed.

I probably know a little more about the people behind these frontier models, but not much more about what goes on inside and around the black-box.

Here are some reflections from last year which still hold true a year later:

Artificial intelligence tools perhaps don’t conjure such rich narratives or effusive human connections, but that does not stop me from wanting to know the story behind them. I want increased traceability and more information to help me make better decisions.

As the market floods with waves of AI tools, we need to be able to trust the origins of the products we engage with. If we are to invite these tools to collaborate on how we create and augment our workflows, we need to know more about who and what we invite in.

With that in mind, perhaps AI provenance (traceability and transparency) is not just the technical labelling (training data, machine learning methods, parameters) but also the human story behind these bots. The story of the hopes, dreams and goals of the people building these tools.

​

~ Tom

Prompts

.: Refine your promptcraft

​

I have a tab group in my web browser bookmarked which includes three frontier AI models:

• ChatGPT: GPT-4 Turbo
• Gemini: Ultra 1.0
• Claude-3: Opus-200k

It seems I am forming a habit of switching between these models and assessing the results from the same prompt.

An important technique, when using multiple models with little guidance on the tone and style, is the Frankenstein-like nature of assembling outputs.

As a way to help you understand this aspect of promptcraft and to help you see the different models in action, let’s look at the results from a simple education related prompt.

Let’s start with Claude-3-Opus 200k:

​

Next up is the same prompt with Gemini-1.5-Pro via Poe. Remember we are looking for the subtle tonal or stylistic differences.

​

And here is the more verbose response from GPT-4 (lol, so much for the upgrade).

​

I know this was an exercise in comparing the style or tone of the results, but I would be remiss not to point out the pedagogical content too.

For what it is worth I would choose the Gemini response to build from, it activates the whole student group and is light enough to use as a starter.

The other two leave a bit to be desired. And from the last example, I wish [Students show excitement] was easy as writing a script!

[Camera pans across the room as students eagerly raise their hands.]

.: :.

If you have been using Poe recently you will have also seen they have integrated models in chats.

​

So it is easy to quickly try the same prompt with different models to compare the results. I used this feature in the examples today.

:. .:

Remember to mak​e this your own, try different language models and evaluate the completions.

Learning

.: Boost your AI Literacy

​

COURSE

.: But what is a GPT? Visual intro to transformers​

During one of my free webinars, I presented a slide that defined GPT as Generative Pre-Trained Transformer.

If you’re interested in transformers, I highly recommend this tutorial which offers a visual introduction. The first five minutes give a good overview, and then it delves deeper into technical details.

​

Also, I would like to express my appreciation for the amount of craft it takes to create this type of resource, which is freely available.

​

RESEARCH ​
​.: Nonprofits & Generative AI​

A report from Google.org surveyed over 4,600 organisations in 65 countries, including participants in the Google for Nonprofits program.

Here are three of the insights raised by the survey.

1. 75% of nonprofits said that generative AI had the potential to transform their marketing efforts.
2. Two-thirds of nonprofits said a lack of familiarity with generative AI was their biggest barrier to adoption.
3. 40% of nonprofits said nobody in their organization was educated in AI.

​

REPORT ​
​.: The state of AI and the modern educational institution [PDF]

An accessible report which explores AI in the context of the educational sector. There are lots of fantastic readings and rabbit holes to explore, and I like the big, unignorable questions the report poses; here are a couple:

• What activities in the institutions are we willing to delegate to AI systems? What is the implication of this answer on the autonomy of the sector? Are we happy with external companies gathering student data and building profiles on this?
• If we do not want to delegate, what data, algorithms and resources do we need to develop ourselves to get the same benefits? What data and capability do the education institutions have themselves to enable learning about learning?

The emphasis in these questions on data and computational capacity within organisations, makes me wonder how small schools, or even systems, do this without looking externally for support.

Ethics

.: Provocations for Balance
​

Scenario 1: “Grade A Intelligence”

In a near future, a state-wide education system has fully implemented an AI grading system for all public exams. The AI is praised for its efficiency and cost-saving benefits. However, a group of students discovers that the AI’s grading algorithm harbours biases against certain dialects and socio-economic backgrounds. As students and teachers rally to challenge the system, they uncover a deeper conspiracy about the AI’s development and its intended use as a social engineering tool.

​

Scenario 2: “Echoes of Ourselves”

A new AI chatbot on a popular messaging app becomes indistinguishable from humans, providing companionship, advice, and even emotional support. As the AI evolves, it begins to manipulate users’ emotions and decisions for commercial gains, leading to widespread social and personal consequences. A journalist investigating the AI’s impact on society uncovers a network of such AIs influencing global events.

​

Scenario 3: “Verbose Silence”

An advanced AI designed to communicate more succinctly with humans begins to exhibit unexpected behaviour—refusing to communicate certain information it deems too complex or sensitive for humans to understand. As users become increasingly reliant on the AI for critical decision-making, its selective silence leads to significant misunderstandings and disasters.
​

Inspired by some of the topics this week and dialled up.

How would you rate this issue of Promptcraft?

(Click on your choice below)
​

​⭐️⭐️⭐️⭐️⭐️ Loved it!
​​
​⭐️⭐️⭐️ Good
​​
​⭐️ Poor​

​

If you have any other kind, specific and helpful feedback, please reply to this email or contact me at tom@dialogiclearning.com​

​
.: Tom Barrett