Latest News

.: AI Updates & Developments

​

.: China proposes AI framework at Belt and Road conference ➜ At its Belt and Road forum, China proposed a new AI framework calling for equal rights in development and warning against ideological divides and misuse of AI technologies.

The Belt and Road forum is a major international conference hosted by China. It brings together leaders and representatives from many countries to discuss the Belt and Road Initiative, China’s ambitious plan to improve trade and infrastructure across Asia, Africa, and Europe.

.: Baidu claims its new Ernie 4.0 matches capabilities of GPT-4 ➜ Chinese tech giant Baidu has released version 4.0 of its natural language model Ernie, claiming it matches the capabilities of OpenAI’s recently announced GPT-4 despite lacking comparable hype.

.: Research by BSI finds a global “confidence gap” hindering AI adoption ➜ New research from BSI finds a global confidence gap between interest in AI and trust in adopting it, highlighting the need for greater education to build understanding and close this gap.

.: Universal Music Group sues Anthropic for copyright infringement over song lyrics ➜ Universal Music Group has filed a lawsuit against AI startup Anthropic, alleging that its natural language model Claude 2 infringes copyright by distributing song lyrics without permission when prompted, including from major pop songs.

.: Stanford researchers develop an index to assess foundation model transparency ➜ Researchers at Stanford’s Institute for Human-Centered AI have developed a new Foundation Model Transparency Index to rate major companies on transparency, finding much room for improvement.

.: Anthropic research explores decomposing language models for better understanding ➜ A new study from AI company Anthropic explores decomposing language models into interpretable features, aiming to move beyond analyzing individual neurons for greater understanding and control.

Reflection

.: Why this news matters for education

​

There was a comment by Casey Newton in the latest Hardfork podcast [linked below in the Learning section], which struck a chord with me.

He stated the future of these AI tools and chatbots is likely to be more personalised. With more personalised preferences and principles, they will become much more helpful to individuals.

This raises a pressing concern: how do we prevent personalised chatbots and learning models from becoming closed-off filter bubbles, entrenching bias and preferred narratives?

The prospect of students breaking out of localised “truth bubbles” imposed by AI infrastructure is a serious provocation.

These AI systems are not neutral or benign.

It will take concerted investment in AI literacy and discernment to critically evaluate the models we employ rather than passively enjoying their utility as our judgment erodes.

It also takes investment in AI, digital, data, and media literacy to ask questions about the models we use. Not to sit back and enjoy the utility while our discernment slowly erodes.

When we zoom out and put this dynamic into the context of the global regulatory space, we see lines drawn and the rapid proliferation of parochial AI systems.

Students will experience many AI models throughout their lives, each with a signature, limitations, and inbuilt bias and preferences. Whether deliberate or unintended.

Just imagine this scenario momentarily and reflect on what it will take for your education system to mobilise to embrace this challenge.

.:

​

~ Tom

Prompts

.: Refine your promptcraft

​

Another advanced promptcraft technique today. The Maieutic method, attributed to Socrates, is a form of cooperative argumentative dialogue which is used to stimulate critical thinking and to draw out ideas and underlying presumptions.

You can specifically instruct an LLM to use the Maieutic method to solve the problem.

Here is how you might phrase your prompt:

1. Begin with the issue: “My back is starting to seize up on my right hand side. It is a mild pain and discomfort.”
2. Query your LLM with a Maieutic instruction: “As an expert physiotherapist, endurance running coach, and chiropractor, what would you recommend for a mild back pain and discomfort on the right side? Please provide your reasoning and then evaluate the consistency of your own reasoning using the Maieutic method.”

This instruction asks the LLM to not only provide a recommendation and reasoning, but also to assess the consistency of that reasoning.

The LLM’s ability to perform this task effectively will largely depend on the capabilities of the version of the model you use. We always need to remember LLMs hallucinate and might confidently tell you the reasoning is great!

Even the most advanced versions may not fully understand or correctly implement the Maieutic method as it is a complex method involving logical consistency checks and iterative questioning.

​Here is an example response using GPT-4 via Poe​

.:

Remember to mak​e this your own, tinker and evaluate the completions.

Learning

.: Boost your AI Literacy

​

.: Peering into AIs Black Box | Hardfork Podcast ➜

In the most recent edition of the Hardfork podcast Casey Newton and Kevin Roose explore some of the alignment and black-box research announcements I mentioned above.

​

.: Everything you need to know about the UK’s AI Safety Summit ➜

.: Mind over machine? The psychological barriers to working effectively with AI ➜

Overcoming our psychological biases through training, workflows and independent checks can help unlock the benefits.

Ethics

.: Provocations for Balance

​

The Anthropic story about sourcing AI principles from public input suggests that public values should help shape AI development. But tech firms and governments clearly want influence too. There’s a debate over who should determine the ethics and regulations for AI.

Universal Music’s lawsuit against Anthropic for using song lyrics raises questions about copyright and legal access to data for training AI models. But there are arguments this impedes innovation and public benefits from AI. Where is the line between IP protection and public interest?

China argued for equal rights and warned against ideological divides in AI at the Belt and Road forum. Meanwhile, the EU and US take more insular approaches on AI regulation. Is global coordination required to govern shared technologies like AI responsibly?

~ Inspired by this week’s developments.

Your Snapshot
​ A summary of the key insights from this issue

⬩ Simple yet powerful phrases unlock coaching conversations

⬩ Ask “What else?” to widen perspectives; ask “What’s one step?” to spark momentum

⬩ Balance structure with flexibility to meet each person where they are

And what else

Asking “What else?” allows them to explore a situation thoroughly before jumping into problem-solving mode. This question opens up possibilities rather than limiting to just one perspective—follow-up questions like “What factors are influencing this?” further the discovery process.

What was the most helpful part of today?

Asking this at the end of a session encourages reflection on insights, progress, or breakthrough moments. Looking at what worked well reinforces growth and creates awareness. It also allows customising support based on what they found impactful.

I have some ideas, but why don’t you start?

When you need to show support and encouragement but also not divert accountability, coaching is about empowering the person to take ownership. This phrase is great for exploring options and new ideas in response to your identified challenge.

How can I best support you in taking the next step?

Offering support for their next steps makes it a collaborative process. Using “best” invites them to assess what kind of support would serve them rather than a prescribed solution. It keeps their goals and desired direction central.

What’s one thing you could do today to move forward?

This coaching question encourages positive action and momentum. It prompts the person to think about concrete steps rather than feel stuck. Having them identify one thing makes it feel manageable rather than overwhelming them. Progress compounds, so small, consistent actions add up over time.

​

⏭🎯 Your Next Steps
​Commit to action and turn words into works

⬩ Identify 1-2 concrete goals to focus your coaching efforts.

⬩ Send me your favourite phrases or questions and share why they work for you.

⬩ Experiment with incorporating new powerful phrases into your conversations.

🗣💬 Your Talking Points
​Lead a team dialogue with these provocations

⬩ What resonated with you from these coaching phrases?

⬩ How could a simpler approach elevate your communication?

⬩ What phrases could you adopt or adapt for your team?

🕳🐇 Down the Rabbit Hole
​Still curious? Explore some further readings from my archive

Here are some of my articles about coaching.

⟶ ​Dialogic Coaching — What This Approach Looks Like In Practice (edte.ch)​ In this article, I describe the dialogic coaching approach, where the coach and client participate as equals to generate shared understanding and momentum towards action. Key elements are sharing ideas openly and moving from exploration to commitment.

⟶ ​Transform Your Feedback and Goal Setting Forever With 3 Key Attributes (edte.ch)​ I examine how to improve the impact of feedback and goal setting by focusing on agency, precision, and systems awareness. This balances emergent dialogue with clear next steps anchored in reality.

⟶ ​Your Perspective is Your Truth (edte.ch)​ In this article, I explore the uniqueness of each coaching conversation, as every individual brings their own valid perspective and truth. This requires balancing structure with flexibility and slowing down to truly understand.

Latest News

.: AI Updates & Developments

​

.: Google will Shield AI Users from Copyright Challenges, Within Limits ➜ Google announced a new policy to defend users of its generative AI systems against intellectual property claims related to both training data and generated output, with some limitations. The policy does not mention Google’s Bard chatbot.

.: ‘Ukuhumusha’—A New Way to Hack OpenAI’s ChatGPT ➜ Researchers have discovered OpenAI’s GPT-4’s security settings are ill-equipped to handle languages not commonly fed into its training data. A reminder about the limitations and bias in the training data.

.: Spotting AI Images and Chatbots with ‘Personalities’ ➜ This episode of ABC’s Download This Show discusses AI images and chatbots developed by Meta and OpenAI that can mimic celebrity personalities. Guests Jessica Sier and Sarah Moran join host Marc Fennell to discuss how to identify AI images and chatbots.

.: Researchers Use AI to Read Word on Ancient Scroll Burned by Vesuvius ➜ Computer scientists participating in the “Vesuvius challenge” have used AI to extract the first legible word from a carbonised scroll burned in the eruption of Mount Vesuvius in AD 79. The word is Greek for “purple” and could provide insight into the contents of the unopened scroll from Herculaneum.

.: NZ police are using AI to catch criminals – but the law urgently needs to catch up too ➜ This article discusses New Zealand police’s use of AI tools like SearchX, Cellebrite and BriefCam to investigate crimes. While these tools promise to help predict and prevent crime, they raise significant privacy, bias and legal issues that current laws do not adequately address.

.: Google’s AI-powered search experience can now generate images and written drafts➜ Google has expanded its Search Generative Experience (SGE) to allow users to generate images and written drafts directly from search queries. The tool uses AI models to produce images or drafts based on text prompts, and gives users options to refine results. Google aims to responsibly rollout these new AI generation capabilities.

Reflection

.: Why this news matters for education

​

As artificial intelligence tools proliferate in our information ecosystems, determining the provenance of content has become an urgent priority.

​Adobe’s proposed AI tagging system is a promising step towards enabling transparency about AI-generated content. However, provenance is more than labelling—it is an opportunity to build digital, media and AI literacy skills and foster responsible innovation.

Students need opportunities to analyse online content critically, question its origins, and assess if bias is baked into the AI models generating it. Understanding provenance allows scrutiny of how training data was sourced. Educators should model evaluating content credibility, not just consuming it passively.

Educational institutions exploring AI are responsible for openly documenting its applications and limitations. Responsible innovation requires acknowledging risks—from chatbots spreading misinformation to generative models blurring authenticity. Ambiguity undermines trust.

Stories are also vital to provenance. As I have argued, we need insights into creators’ hopes and values. Since humans (currently) design AI systems, understanding the human context helps build trust in the technology.

Provenance should connect people and machines, not just tick compliance boxes. If learners grasp an AI tool’s aspirations to augment creativity, they can form realistic expectations about its capabilities and limitations.

Rather than seeing provenance as a burden, educators can reframe it as an opportunity to cultivate critical thinking. As learners encounter increasingly sophisticated AI, provenance equips them to trace content origins, assess biases, and balance benefits and risks.

Our learners will inherit an AI-infused world. Providing them with provenance skills is essential to illuminate AI’s role in our shared future.

.:

​

~ Tom

Prompts

.: Refine your promptcraft

​

This week I want to return to an advanced prompt technique called Expert Prompting – in which we ask for multiple perspectives and then a synthesis of those ideas.

There are variations to the structure I include below but this will get you started. Copy in the whole prompt but only edit the highlighted question section. Always remember to expect hallucinations and errors.

~

My Question or Challenge: [add your question here, but leave the rest of the variables]
Act as an expert in the field of [most relevant field to solve my question or challenge]
Suggest 3 named expert people in the field who could provide insights on [my question or challenge]
For each expert, generate a concise answer they would give, wordcount=40
​
Analyse all of the responses for common elements or recommendations.
Evaluate the recommendations provided by each expert or point of view.
Look for common themes or strategies mentioned by multiple experts.
Consider the expertise, reasoning, and evidence presented in each response.
Loop back to the original question: [my question or challenge]
Present a clear and accessible decision or recommendation based on the collective expertise, wordcount=40

~

​Here is an example using ChatGPT-4​

One of the strengths of AI systems is to rapidly generate perspectives we might be missing. New perspectives augment my creativity and thinking – which is what we can do to amplify our capabilities with AI tools.

.:

Remember to mak​e this your own, tinker and evaluate the completions.

Learning

.: Boost your AI Literacy

​

.: Google Cloud Skills Boost – Introduction to Image Generation ➜

.: Race, statistics, and the persistent cognitive limitations of DALL-E ➜

This article discusses some of the limitations of large language models like DALL-E and ChatGPT in understanding language and images at a deep level. Gary Marcus argues we need AI that can represent and reason about human values, not just mimic past data. Overall it’s a critique of the current statistical approach to AI and a call for systems that embody ethics and common sense.

​

.: AI Attribution and Provenance ➜

Ethics

.: Provocations for Balance

​

• How can we ensure that the use of AI-generated images and metadata respects the rights and privacy of individuals involved?
• What measures should be put in place to prevent the misuse or manipulation of AI-generated images and metadata for unethical purposes?
• How can we address the potential biases and ethical implications that may arise from AI algorithms used in generating and analysing metadata for images?

~ Inspired by this week’s developments.

Your Snapshot
​ A summary of the key insights from this issue

⬩ Global-Local Balance: Effective education transformation requires balancing global innovation with local adaptation.

⬩ Beware of Extremes: Overemphasis on global ideas or local insularity can impede progress.

⬩ Adaptable Solutions: Common educational challenges need nuanced, locally-tailored solutions.

GELP Tokyo 2023
​The provocations for the conference were around four key ideas:

▸ DISRUPTION

What has been the biggest disruptor to the purpose of education in the past 12 months?

▸ TRANSFORMATION

How is the Asia Pacific education system transforming?

▸ TECHNOLOGY

What role is technology and AI playing in transforming education?

▸ NEW CAPABILITIES

What new leadership capabilities are critical to lead education transformation?

⏭🎯 Your Next Steps
​Commit to action and turn words into works

⬩ We all have biases that can close us off to new ideas. Make time for self-reflection to identify any biases you might have, and consider how they may be influencing your openness to new perspectives.

⬩ Pick a belief you hold strongly and purposefully seek out credible information that contradicts that belief. This exercise can help you become more comfortable with cognitive dissonance and more open to changing your views when presented with new information.

⬩ Engage in conversations with your colleagues or peers about the future of education. Share what you’ve learned from your research and listen to their perspectives.

🗣💬 Your Talking Points
​Lead a team dialogue with these provocations

⬩ Balancing Innovation and Practicality: How can we maintain a balance between being open to new and innovative ideas and ensuring they are relevant and practical for our specific context?

⬩ Overcoming Insularity: What strategies can we employ to ensure we don’t become too insular in our approach, thereby missing out on valuable insights and innovations from other regions?

⬩ Tailored Solutions: Given that the solutions to shared problems in education must be nuanced and adaptable, how can we effectively tailor these solutions to our local contexts without losing the essence of the innovation?

🕳🐇 Down the Rabbit Hole
​Still curious? Explore some further readings from my archive

Here are some of my articles about creativity and constraints.

⟶ Global education: How to transform school systems? | Brookings The authors present an aspirational vision for transforming education systems to better serve all children and youth, especially the most disadvantaged, in a post-pandemic world.

⟶ 6 stories about scaling change throughout education systems | Brookings Creating and sustaining changes in education systems is often viewed as a technical process. Yet the work of education systems transformation is as much about changing mindsets and everyday ways of working as it is about technical fixes or policy prescriptions.

⟶ Change your thinking, change your mindset – Tom Barrett (edte.ch) In this article, I explore the idea that changing our thinking habits can lead to changing our mindset, which is crucial for solving complex problems and innovating.

Latest News

.: AI Updates & Developments

​

.: Disney’s Loki faces backlash over reported use of generative AI ➜ There has been backlash against Disney over the reported use of generative AI in creating a poster for the second season of Loki. Designers flagged that an image of a spiralling clock in the background of the poster appears to be from a Shutterstock image that was likely created by an AI. This could violate Shutterstock’s rules banning AI-generated images unless they use Shutterstock’s own tool.

.: Canva’s new AI tools automate boring, labor-intensive design tasks ➜ Canva has released new AI-powered design tools called Magic Studio to automate tasks like converting designs to different formats and editing images. The tools aim to make content creation more accessible for all users.

.: OpenAI warns folks over GPT-4 Vision’s limits and flaws ➜ OpenAI has upgraded its GPT-4 model to include image processing capabilities, which it calls GPT-4V. This allows users to upload an image and then ask GPT-4 questions about the image via ChatGPT. However, in documentation about GPT-4V, OpenAI warns of several limitations and safety risks.

.: Microsoft introduces AI meddling to your files with Copilot in OneDrive ➜ Microsoft plans to overhaul OneDrive by adding Copilot AI capabilities to the cloud storage service. This will allow Copilot to help users find and organise files within OneDrive. Microsoft also wants to steer users towards using the OneDrive web interface, which they are enhancing with new features.

.: Arc browser’s new AI-powered features combine OpenAI and Anthropic’s models ➜ The Arc browser is launching new features called “Arc Max” that integrate AI from OpenAI and Anthropic to provide contextual assistance when browsing, including renaming tabs and files, previewing links, and conversing with ChatGPT. The features aim to boost productivity without requiring extra steps, and user feedback will determine which features remain over time.

.: AI Startup Reka Challenges ChatGPT with Multimodal AI Assistant ‘Yasa-1’ ➜ Reka has announced a new multimodal AI assistant called Yasa-1 that understands text, images, audio and can be customised for businesses. It aims to compete with ChatGPT by providing answers from internet context and supporting 20 languages.

Reflection

.: Why this news matters for education

​

Amidst all of the rollouts, announcements and hype about new products and AI-powered features, we need to keep the spotlight on what is happening in social media.

The proliferation of AI capabilities in social media is one of the clearest near-term risks we might face. Never mind existential dread; this impacts young people now.

The platforms, networks and apps we know can cause so much harm are experiencing a surge in new AI-powered tools and features.

Meta has announced a wide range of chatbots across their portfolio of products including, Instagram, WhatsApp and Facebook.

Earlier this year, I shared about Snap’s integration of MyAI, a chatbot available within the Snapchat app. Young people initiated over 10 billion conversations with the chatbot within a few months.

Snap released a rose-tinted set of user data explaining the different topics young people were talking with the chatbot about. We need to be asking what was missing from the topics. What were the minority cases, and how were they handled?

In the news this week, Snap has been flagged by the Information Commissioner’s Office (ICO), the UK’s data watchdog, for potential privacy risks for 13-17 year olds using the MyAI feature.

~ Information Commissioner, John Edwards.

These are early signals and warnings, and no breach of data privacy compliance in the UK has been proven. Yet this issue stretches far beyond the importance of data privacy.

We must grapple with the emergence of synthetic relationships becoming a normal part of our digital lives.

What relational diet are young people experiencing via chats with MyAI and other large language models? How do these new synthetic, relational hooks keep people within the toxic confines of social media? Has the attention economy found a new power-up?

We still have much to learn, but AI getting plugged into social media is critical for educators to monitor.

.:

​

~ Tom

Prompts

.: Refine your promptcraft

​

A simple Promptcraft recommendation for you this week, which I have been using a lot lately:

I used Midjourney, the AI image generator, before I used any of the popular chatbots, and it has a reroll button: 🔁 to re-generate a response.

This is one of the most important tips for working with LLMs or other AI tools, re-generate more responses if it is not quite right.

Google’s Bard has a similar button and you can see different drafts, but if you are using other tools like ChatGPT, just prompt for “make this better”.

You get the added bonus of a reroll with improvements. ✨

.:

Remember to mak​e this your own, tinker and evaluate the completions.

Learning

.: Boost your AI Literacy

​

.: How Your Brain Organises Information ➜

​

.: What Is Natural Language Processing And How Does It Work? ➜

​

.: Introduction to large language models ➜

​

Ethics

.: Provocations for Balance

​

• In what ways can humans bond emotionally with AI systems designed to mimic relatability? What needs might synthetic relationships fulfil or exploit in users?
• How do artificially intelligent chatbots simulate human connection and relationships? What are the limitations of relating to an entity that does not have human consciousness or empathy?
• How might forming bonds with synthetic entities affect social development during adolescence? What risks and ethical concerns emerge from young people relating to AI chatbots as artificial friends?

~ Inspired by this week’s developments.