Generative AI and DCF Proposals - Letter to the Welsh Government's Education Minister

Correspondence.Anna.Brychan@gov.wales

August 11th, 2026 

Dear Anna Brychan MS.

I hope you don't mind me contacting you, but after 30 years of being involved in education in Wales, this is the first time I have urgently felt the need to share my thoughts with a minister. Before I outline some of my concerns and proposals, I'll quickly let you know who I am. My name is Gareth Morgan and I am an independent educational digital technology consultant, primarily working with primary schools across Wales. I started my teaching career in 1996 before becoming an ICT curriculum improvement advisor for Newport School Improvement Services. For many years I was a NAACE ICT Mark assessor and in 2012 was an advisory member of the Welsh Government task and finish group who wrote the "Find it, make it, use it, share it: learning in digital Wales" document, which helped to shape the way forward for the use of digital technologies by schools, practitioners and learners across Wales. For the past 14 years, I've continued to work with schools through my own consultancy where I provide advice, support and training to primary school practitioners from all across Wales, focused around the Digital Competence Framework (DCF) and the computation element from the Science and Technology AoLE. I have therefore spent nearly all my teaching career heavily involved in supporting the use of digital technologies across the curriculum. Recently I have been reflecting on what has happened in the world of 'EdTech' over the last 30 years and also spending much time focusing on the future, especially with regards to the recent hype around generative AI. 

From what I have gathered from colleagues and after attending a ‘DCF national network conversation’ meeting in Wrexham last July, my understanding is that the DCF is going to be updated shortly. Therefore this seems like the perfect opportunity to share with you some proposals outlined below which I hope you will consider.

Proposal 1 
A minimum 3 year moratorium on the use of generative AI by learners. During this period the DCF is updated to include reference to 'unplugged' critical AI literacy. 

Setting the scene 

I, along with many others within education, and increasingly parents and organisations, are concerned about the use of generative AI products by our learners across all sectors of education. Justifiably many have ethical concerns based around the impact that this technology is already having on our society, on the environment and on human intelligence. As a teacher and advisor, I am particularly concerned about introducing to our learners an untested digital technology that many are already highlighting, can erode critical thinking and their ability to problem solve. I believe that this moratorium would provide the whole education community the time to reflect on evidence of whether generative AI may or may not have a positive impact on learner standards. This evidence has to be soundly based upon rigorous, independent, academic research. The period should also provide the space for learners, educators and the wider school community, to ask questions and to debate whether or not this is even a technology that should have a place within education. 

Racing to implement AI and the importance of learning in the primary school 

I propose, at the minimum, a 3 year moratorium on the use of any generative AI product by all our learners in the classroom. I am very concerned that we are racing into its introduction into schools, especially if an updated DCF carries descriptions of learning relating to AI proficiency, for example, "I can prompt a chatbot to answer or 'create' an output." I worry about the consequences on our learners. Primary school in particular is an extremely important fundamental stage in child development and early learning, and any introduction of new digital products has to be thoroughly scrutinised before its introduction. Our young learners should not, as I see it, be guinea pigs in a big tech experiment; we should not be trialing platforms to see what might be possible. An example of which I saw recently published by Estyn. This should be the job of the AI developers, not educators. Learners only have one chance in the primary school and we should not let an experiment with generative AI technology hinder an extremely important part in learner development, without being 100% certain that what we are introducing is beneficial to the learner. If the evidence or the data does not exist, then generative AI is not a viable product and does not belong in our classrooms. To borrow and adapt someone else's quote, "when digital technology fails, the cost isn't measured in pounds, it's measured in the learning and development of our children."

The generative AI hype cycle and a history of digital technology failures in education - "Technology is the revolution still coming after 50 years." (Dylan Wiliam) 

Read any news articles or marketing literature about generative AI and you will soon come across the word 'potential'. It has been a reoccurring word throughout my career whenever a tech company wants to introduce its latest device or application into education. Unfortunately, only to see a couple of years later that potential and promises have yet again failed to deliver any measurable benefits to our learners. Take for example the Welsh Government's initiative to introduce interactive whiteboards into the classroom in the early 2000s or the school led drive to introduce Apple iPad devices into the classroom in the mid 2010s. Examples where there was enormous hype surrounding these technologies, with much being promised by companies and educationalists about how these devices would have the potential to revolutionise the classroom and improve learning standards. Only to see after a couple of years them failing spectacularly on improving learning and teaching. My worry is that education is not learning from the history of these and other failures, and that this latest generative AI hype cycle, which appears to be fully embraced by both UK and the previous Welsh Government, could well prove to be another such example. The introduction of any new device or application affects both the learner and the teacher over the period of the cycle. There are costs to the school, both financial (purchases, staff training) and a cost to learning, where learners may have been exposed to teaching and learning strategies based around this new tech that at best may not move learning on and at worst is detrimental to the learning process. Generative AI looks to be another example of a new digital technology dropped into education, with educators and learners left to work out what to do with it. Like so many others, this is a digital technology product that wasn't asked for by the education community, attempting to solve a problem nobody actually had. Let us make sure that during this 3 year pause we make certain that what we may introduce to our learners is actually helping them and not hindering their progress. 

Use the moratorium to scrutinise rigorous, independent, academic research and for all education stakeholders to be kept fully abreast of what this research is finding and to be able to ask questions. 

This 3 year break will provide time for the Welsh Government, Estyn, consortia, local authorities, schools and the wider school community to ask questions and to gain a better understanding of the implications to learners and teachers of using generative AI in the classroom. This period will also give the education community the breathing space to evaluate independent, rigorous, academic studies on the impact of generative AI on learning standards. As was pointed out in a recent report produced by England's Department for Education, personalised learning technologies, especially those which integrate generative AI, found that "there is still limited evidence regarding their long-term impact on learning outcomes." I am hopeful that over the period of the moratorium more studies will be produced and if these find there is none or limited impact on learning, then schools would not have wasted their valuable teaching time dabbling with generative AI. However, if studies do find a positive impact on learning outcomes, then share the results with schools, but with the caveat that any introduction into the classroom has to be closely based within the parameters of the successful academic study.

Switch off all generative AI features on Hwb platform applications. These products are not designed for young learners.

Currently Microsoft 365, Google Workspace for Education and Adobe Express (all Hwb hosted applications) are accessible by our learners. Each of these products has generative AI features that are already switched on by default or built in for learners to use, whether schools, learners or parents wanted it or not. This I would argue is questionable. Teachers and parents should be able to retain control over the products that their learners or children have access to. For some that would include not allowing learners to access to generative AI. During the moratorium I would like to see these features disabled for our learners. If individual platforms are unable to do this, then it should be recommended to schools that they do not use these products with their pupils until there is clear evidence of the benefits to learning at the end of the moratorium. It is also worth noting that as the Alan Turing Institute recently pointed out, each of these platforms and their generative AI products were never designed for young children's learning or with safeguarding in mind, risking their exposure to potentially harmful or inappropriate content. The urgent need for critical AI literacy, NOT AI proficiency I partially agree with Estyn's recently published report and recommendations into how AI is currently supporting teaching and learning. I certainly believe that an updated DCF should include "critical evaluation" and "ethical understanding", but I disagree with the suggestion of "AI-related digital literacy," until the end of the moratorium at the very least. During this period our focus should be on Critical AI literacy, which is not about training learners on how to prompt generative AI products, which is AI proficiency or AI literacy as Estyn puts it. Critical AI literacy should open the learner's eyes to the social, intellectual and environmental impact of the introduction and use of these products. We could introduce these issues through 'unplugged 'activities without the need for learners to actually use them, in much the same way as many of the Citizenship descriptions of learning are currently taught within the DCF. Think for a moment about how we teach 10 or 11 year old learners an awareness of the dangers associated with social media, smoking, drug use or alcohol. In each of these examples, we highlight and discuss the issues but we never ask learners in the classroom to actually login to a social network, smoke a cigarette or drink a beer. 

Critical AI literacy is about fully informing our learners, with the aim for them to be able to eventually make an informed decision about when, and importantly, when they shouldn't use generative AI in their learning. At the heart of the Curriculum for Wales are the four purposes and if we want our learners to be "ethical and informed" citizens, then we need to provide them with the knowledge about generative AI to be able to make that informed decision. A learner who understands why the prompt produces what it produces, who built the system and why, and what it costs them cognitively and to the planet environmentally, has developed critical AI literacy. Therefore this must urgently be included in a new DCF update.

Is there a hypocrisy in the Welsh Government's encouragement in the use of generative AI by education and their commitment to net zero and sustainable development? 

The Welsh Government often presents itself as a global leader when it comes to its commitment to net zero and sustainability policies. I believe there is a tension between the Welsh Government’s green ambitions and its drive for public and private sector generative AI adoption. For example, it is proposed that five new hyper-scale data centres are going to be built across Wales. One of them, in Newport, is being built for Microsoft. The very company that provides Office 365 and Co-Pilot to our learners through the Hwb platform. Yet this is a company that has seen global carbon emissions rising by 29% over the last four or five years and is predicted to rise for several more years. As well as these data centres consuming huge amounts of energy, fresh drinking water is often used to cool the power hungry chips. Water UK, who represent the water companies have recently reported that the UK will not have enough water to support this predicted growth. With this as a background, I would argue that it is difficult for the Welsh Government to highlight their green credentials while also encouraging the growth of hyper-scale data centres across Wales and the generative AI use in education (and other sectors). Increasing the use of generative AI can only exacerbate environmental problems caused by these data centres. There also seems to be a gap between green policy aspirations and the technical reality. For example, while public sector organisations like Estyn, advocate for ethical, sustainable AI governance, their guidance actually clashes with how hyper-scale cloud computing and AI platforms operate. According to Estyn's website, they have an 'ambition' to "utilise data centres using renewable energy sources" when using generative AI. This is certainly 'ambitious', if not a contentious statement. A similar statement is also suggested in a Hwb document to schools about generative AI in education. According to the document it states that it is essential for schools "to adopt energy-efficient AI systems and ensure their use is both strategic and sustainable." From my understanding, it is virtually impossible for an individual school or Estyn to know exactly what an "energy-efficient AI system" is or which physical data centre is handling a specific generative AI request. For example, if a teacher or learner uses Microsoft Copilot, which they would be encouraged to use through the Hwb platform, the query does not travel to a single, static data centre. The prompt could be processed in a number of different data centres across the UK and beyond. If the prompt can be processed in any number of data centres, a school or Estyn what not know if it was going to a data centre that happened to be using renewable energy sources. 

Wales would not be alone in this more reflective, inclusive and critical approach to our learners use of generative AI. Norway has recently imposed a near ban on the use of generative AI products by learners aged 6 to 13 years old, with the Norwegian Prime Minister worried about these tools creating a "short-cut trap" where learners skip the vital stages of independent thinking. Their government also stressing a renewed focus on reading, writing, and mathematics. 

Many teacher unions and organisations are also calling for similar moratoriums or roll backs on the use of generative AI in schools too. In the US, Fairplay, a massive coalition of more than 250 medical experts, child development researchers, and education organisations is calling for a five year moratorium on all student facing generative AI products across K-12 in the US and Canada. In the UK, a coalition of nine teaching unions issued a joint declaration demanding a halt to unilateral government and corporate rollout of classroom AI. Calling for educators and their unions to be involved at every stage of any AI adoption in education and that "AI in education is shaped by professionalism and sound pedagogy."

Proposal 2 
The removal of the PS1 and PS2 statements of what matters from the DCF and from the computation aspect of the Science and Technology AoLE. The slow introduction of digital proficiency, critical digital literacy and critical AI literacy skills from PS3 and beyond.

Setting the scene 

Before I begin outlining my reasons why I believe Wales should consider this proposal, let us look at it in a wider context. Again Wales would not be alone if they acted on this recommendation. There does appear to be a wide movement calling for or implementing similar proposals. Governments and organisations are now questioning or reappraising their policies on the learner use of digital technologies in the classroom and outside it, and if they are having the desired positive impact on our young people. We are now seeing social media bans for under 16s and the banning of mobile phone usage in schools. Several countries, including Sweden, Denmark and Finland, are currently shifting away from their previous classroom heavy tech policies and rolling back the use of digital technologies. Many are now refocusing on the core purposes of the primary school, removing or minimising the use of digital technologies by its learners, especially younger learners, with the belief that this will begin to help them address falling literacy and numeracy standards and also improving their PISA standings, which have seen a decline. Through the removal of the technologies they hope to create better conditions for children to concentrate and develop their reading, writing, communication, mathematical and social skills. Countries that have invested heavily in digital technologies for their schools are now moving some of that investment into more 'traditional' classroom learning materials such as physical books and improving school libraries. 

Therefore with the proposed update to the DCF on the horizon, I think that this is the perfect opportunity for a new Welsh Government to honestly reflect on the impact of many years of considerable investment in digital technologies in our schools. Much of that investment going into the Hwb platform and 1:1 devices in many classrooms which are used by schools to support the implementation of the DCF. I question whether the huge investment in the online technologies, digital devices and the framework are ultimately having the fundamental outcomes we all want for all our learners, especially within the primary sector.

Child development 

Early years educators globally report that there has been a severe decline in 'school readiness' among incoming learners. Some studies have shown that up to a third of children entering reception or nursery lack basic developmental milestones, which can create many challenges for schools. Alongside a growing proportion of children arriving without being toilet trained, unable to dress themselves or with speech and language delays, educators are linking many behavioural problems to heavy exposure to screens and the use of digital technologies by adults, as 'emotion pacifiers'. Educators are reporting that the increased exposure to screens is now resulting in several issues. 

● An erosion in children's self-regulation skills. Where digital devices are regularly handed to very young children to calm their tantrums or prevent boredom they miss crucial opportunities to develop internal emotional regulation. 

● A literacy and tactile disconnect. Where children do not know how to handle physical books and will 'tap and swipe' paper. 

● Fine motor skills problems. Caused by too much screen use which displaces tactile play, such as, block building or scribbling, which leaves children with weaker grip strength and poor hand-eye coordination for pencil holding. Also, a child cannot learn spatial mechanics by moving a finger across a flat piece of glass. 

● Parental 'technoference'. Where excessive adult screen time is limiting the natural, passive language modelling and eye contact during critical developmental windows. A child's language acquisition and emotional intelligence rely entirely on rapid, face-to-face verbal and facial exchanges with humans.

Therefore, this combination of a lack of physical play, reduced shared reading at home, and the emotional blanket via screens, is arguably leaving our educators spending hours each day on basic developmental support before they can even focus on foundational learning. In my opinion, it would then seem ill advised to begin introducing screens into the PS1 classroom that would actually undermine what the school is putting into place, as being a space to develop vital face-to-face socialisation, active play, and language acquisition. Especially as in some of these cases, it was screen usage by children that was at the root of many of the problems. 

I don't think that many would disagree that PS1 and PS2 classes lay the structural, social, and academic foundations for a learner's entire education. In these classes, learning is inherently tactile, physical, and social and it is where educators build the key foundational skills. However, as I outlined above, this approach is becoming increasingly challenging for early education. Therefore, asking a young learner to save files to a specific location; to login to collaborative platforms; or creating, editing and organising multimedia components, (as the DCF currently requires) arguably serve no vital developmental purpose. In my mind this is purely an unwanted and unneeded distraction. Removing DCF skills at PS1 and PS2 would have little or no impact on the learner's future. These are digital skills that can very easily be introduced in PS3 or beyond. In fact, I believe that mapping adult-centric digital skills backward onto children in the early primary school, who are still mastering fundamental physical, social, and cognitive milestones, is certainly questionable at the very least. 

Reading on-screen makes it harder for learners to process information, handwriting fosters deeper understanding than typing. 

If my proposal was implemented then I believe it would free up more time, to focus on developing foundational literacy and numeracy. Over two decades of research has found that learners comprehend better and retain more when reading from paper than from screens. That handwriting fosters deeper learner understanding and stronger test performance, while also building fine motor skills. Handwriting is slower than typing for some older learners, but pen and paper notetakers are forced to focus on meaning in their notes and this activates thinking in ways that typing cannot. Basically we know that slower writing promotes deeper thinking.

Huge spending by the Welsh Government on digital technologies but no evidence of improvement in learner standards. 

As I pointed out above, governments who have or are considering removing digital technologies from the primary school are very much focusing on falling literacy and numeracy standards in their countries. It is not completely clear from reports whether they are putting the whole blame on learners' use of digital technologies in the classroom. But, I think it is fair to say that they see its use as a distraction to learning and that there appears to be a recognition that after huge investment in digital technologies, they are possibly not seeing any value for money. Wales' last education minister proudly boasted in October 2025 that the Welsh Government had invested £167 million since 2019 in digital skills for schools and at home. But the question seriously needs to be asked, what was the measurable outcome on learning standards from that huge investment of money in digital technology? Based on the most recent PISA results (2022), which I sure you are aware of, reading, maths and science in Wales fell to the lowest levels ever, which was well below the OECD average, ranking us the lowest of the four home countries. Also reports from Estyn consistently highlight literacy and numeracy issues across all education sectors. It is obviously impossible to make a correlation between the £167 million spending on digital skills for home and school and our literacy and numeracy concerns, but I do question what was the purpose of all that spending if it did not improve learner standards in these essential key areas? Where is the accountability for that spend when the Welsh Government does not appear to know whether a learner with a laptop performs better academically than a learner without one? Hwb may provide operational statistics about the number of logins or downloads, for example, but I would say "So, what?" I understand and concede that Hwb has provided cost saving for schools (centralised software purchasing for example), provided a national directory for schools and is helping with digital equity, but these are not academic outcomes, to me this seems like infrastructure. £167 million was spent on the digital infrastructure, but Wales has failed to produce any definitive proof that investing in this infrastructure and having a DCF has made Welsh children more literate, numerate, or academically successful. 

Digital technology - High cost, low to medium learner outcomes 

As I mentioned previously, there is certainly an argument that can be made that our learners use of digital technologies is not value for money. Other than providing a conveyor belt of "future consumers" for Google and Microsoft, a phrase used by Google themselves, evidence has consistently shown that learners use of digital technologies have generally failed to improve standards (there are a couple of very specific examples where this is not the case). The Education Endowment Foundation (EEF) for many years has consistently emphasised that digital technology is not a standalone solution for schools and that any impact depends on how well it is used by educator. A quote that I have heard repeated over many years is "good teaching with digital technologies gets the same results as good teaching without it. It's not the technology that makes the difference, it's the pedagogy," or simply put, it's the pedagogy, not the technology. In my opinion, much of the money that has been invested by the Welsh Government in digital technologies, should have been diverted to quality teacher professional development.

Silicon Valley CEOs send their own children to non-tech focused schools and heavily restrict their children's access to the very technologies they promote to others. 

Montessori and Waldorf schools, which are often used by parents who work for Silicon Valley tech companies, do not introduce digital technology into the primary school. According to the Waldorf School of the Peninsula, which serves many of the tech families, the primary school is "screen free", and that many Silicon Valley families choose the school precisely because they understand technology’s power, and its limits, and "value a school setting that prioritises human development first." Recently, Fortune magazine ran an article titled, "Peter Thiel and other tech billionaires are publicly shielding their children from the products that made them rich". In the article Thiel says that he lets his two young children use screens for only an hour and a half per week. Famously, as far back as 2010, Apple's Steve Jobs told a reporter his kids had never used an iPad and that, “We limit how much technology our kids use at home.” These are parents, who arguably understand the future of tech better than anyone and prioritise schools where their children are engaged in physical movement, knitting, and mud-play over iPads or Chromebooks. I'm sure these schools would agree that the skills learned on a tablet are trivial and easily mastered later, whereas the foundational human focused skills must be hard-wired before the age of 11. 

Why should children in Wales not have similar educational opportunities to the children of big tech bosses?

An update to the DCF (PS3) will be needed reflecting legislation on the banning of social media use for under 16s. 

Along with the introduction of 'unplugged' critical AI literacy as I outlined in my first proposal, the Citizenship strand at PS3 will need updating to reflect potential UK legislation on the banning of social media use by under 16s. 

The general slimming down of content in the DCF at PS3 and a greater focus on critical digital and AI literacy. Focus on unplugged coding the slimming down of some content at PS3 in the Science and Technology AoLE 

There are currently 91 descriptions of learning in DCF across PS1, PS2 and PS3. Removing PS1 and PS2 would narrow the amount to 36 left at PS3. However, I would very much like to see that number reduced further. I would like to see far more emphasis placed on the Citizenship aspect of the DCF, as in my opinion, these statements are the critical statements which are directly related to the learner's lives. For example, ways to protect themselves online, balancing their amount of screen time and appropriate behaviours while online. I would like to see the cutting back in the number of statements in both the Interacting and Collaborating, and Producing strands, simply focusing on making sure that learners can login to platforms such as Hwb and focus on computer application basics. I would also like to see a clearer connection between a digital skill being introduced and a critical focus around that skill (critical digital literacy). For example, a learner is shown how to upload a document to Hwb. There should be a clear link made between the basic skill that has been introduced and to critical digital literacy questions such as "Once we share or save this document online, where does it go? Who is able to see it and is there any information in this document (like a surname, address, or photo) that should stay private?"

PS1 and PS2 descriptions of learning to be removed from the computation strand of Science and Technology AoLE. At PS3 I propose a removal of all descriptions of learning except the three related to coding. The Science and Technology AoLE was introduced in the academic year 2022-23, therefore we are four years into its use in the primary classroom. From my personal experience of working with primary school teachers from a great number of schools across Wales, hardly any teacher understands or is confident in delivering several of the descriptions of learning, for example, how digital devices connect locally or globally, sensors and actuators or how data is stored and processed. I think it is fair to say that the majority of teachers are not 'techy' and asking many of them to address these descriptions I feel is unnecessary. These types of aspects should be introduced in the secondary school, by specialist teachers who understand what those descriptions of learning actually mean. I believe we should concentrate on simple coding which would be 'unplugged' utilising computational thinking type activities. 

Conclusion 

I am not anti-tech, I have actually always considered myself a bit of a 'tech-geek'. What I am though is, pro-learning. I am proud to say that I have throughout my 30 year career I've always made sure that I viewed digital technologies through a critical lens, wanting what is best for our schools, that may or may not have included technology. As increasing evidence mounts with regards to the negative impact generative AI is having on on so many aspects of our lives, with declining literacy and numeracy levels across many countries, with our young learners arriving at school with many issues associated to early introductions to digital technologies and longstanding research highlighting that for all the hundreds of millions of pounds invested in tech over last 30 years, digital technologies rarely, if ever, improve learning standards. I believe that the time is right to contemplate a different approach and actually start learning from our mistakes. 

Apologies for the length of this, but many thanks for taking the time to read. If there is ever an opportunity for me to share my personal thoughts, learning and experiences around these subjects, please feel free to contact me. 

Gareth Morgan, 

Barry, Vale of Glamorgan

email: garethmorgan.consultancy@pm.me