Tag Archives: artificial intelligence

AI and schools: progress since 2018

Posted on by
Reply

Sometimes a visitor to this blog can uncover one of the 1,750 posts that I have completely forgotten about. One such as a post about Artificial Intelligence (AI) that I wrote eight years ago after a Report from a House of Lords Committee AI and education – The view of the House of Lords Committee | John Howson

The DfE at that time were talking about creating some 8,000 new teachers of computing. Isn’t it  interesting how often politicians offer up the promise of large number of new teachers with no funding to allow schools to employ them, even if they could be recruited (think of the present government’s 6,500 promise).

In November 2018, according to the DfE’s Workforce Census, School workforce in England: November 2018 – GOV.UK there were 3,954 teachers of computing working in state secondary schools. Together they taught some 31,190 hours, mostly at Key Stage 3 (3,954) with only 964 hours at Key Stage 5, but 2,289 at Key Stage 4.

Fast forward to the 2025 Workforce Census, and the number of those teaching computing had increased to 7,784. Together they delivered some 64,923 hours of learning. An impressive increase, even against the increase in the number of pupils over the same period.

However, using the DfE’s methodology of counting teachers and subjects, the overall hours taught has changed little between 2018 and 2025, from 3,550,691 in 2018 to 3,553,755 in 2025. Overall teacher numbers teaching across all subjects also remained remarkably statics at 219,269 in 2018 and 219,296 in 2025.

Against these numbers, the increase in teachers of computing, or computer science as it is now termed, is noteworthy, at an increase 3,730, doubling the number of teachers teaching the subject. However, it is still 50% below the 2018 number of 8,000 additional teachers put forward by the DfE.

In my original 2018 post, I also wondered about the moral and ethical issues surrounding AI. I wrote the following:

‘The Committee recognised that the 2014 change to the curriculum on IT in schools across England needed time to take effect. However, the removal of any consideration of moral and ethical issues to do with social media and digital technology from the curriculum was regretted by some witnesses; no doubt more so over recent weeks as the various concerns over social media and the handling of personal data have emerged. Personally, I think the downgrading of Religious Education at examination level, where there was a real opportunity to discuss issues of ethic, morality and philosophy, by excluding the subject from the EBacc was a mistake.’

The hours taught for religious education have increased since then; rising from 117,408 in 2018 to 117,772 in 2025.  The percentage of teachers with a relevant post ‘A’ qualification has also increased from 64% to 81.5% for computer science – as it is now termed – and from 54.1% to 72% for religious education.

However, it is worth noting that at Key Stage 3, where pupils are introduced to subjects, the percentage of qualified teachers in 2025 was still only 76.1% for computer science and even lower at 65.9% for religious education.

It seems the foundations of both subjects in secondary schools are still being taught by those least likely to have a qualification in the subject. Whether or not they are also trained teachers isn’t possible to determine form the simple Teacher Workforce publications.

One wonders what the DfE’s policy for both subjects is in the Curriculum Review. Both are key subjects for our modern age, and I would like to thank the reader that reminded me of my 2018 post, and thus, prompted this one.

Are dashboards useful?

Posted on by
2

LinkedIn invited me to try their new lead generation tool. I asked it to look at this blog, and it came up with a number of possible future posts I could use to gather new readers, all generated by AI (artificial intelligence) from some key prompts and audience identifies that I provided.

It is an interesting piece, although, I have altered (added in italics) some additional text. I would be fascinated to hear what readers think.

AI starts here

A dashboard showing applications up fifteen percent across the board is one of the easier things in education data to misread, and can be one of the more expensive in terms of outcomes.

If you consider recruitment to Initial Teacher Training (ITT) as an example. The structural reason that a dashboard can be misleading is that an ITT recruitment dashboard aggregates things that behave nothing like each other. Domestic and international applicants, candidates with and without visa routes, subjects with entirely different seasonal curves, and applications that will convert at wildly different rates all sit inside the same headline number. The dashboard is honest. The aggregation is what misleads.

Three patterns come up repeatedly:

– Treating a total as a common trend, when a surge in one subject or one applicant group is carrying the whole figure

– Reading application volume as recruitment, when the gap between offer, acceptance and enrolment is where subjects like physics may be very different to that in Classics.

– Comparing the current month to the same month last year without checking whether the policy environment, bursary levels or visa rules changed in between

However, the second-order effect is the one that causes real damage. When a headline figure looks healthy in December, the case for intervention weakens, and by the time the subject-level picture is unambiguous in March the levers that would have mattered, bursary announcements, provider allocations, targeted campaigns, may have already closed for the cycle.

Intervention decisions are can risk being made against an aggregated number at exactly the point in the year when disaggregation would tell you, the reader, something much more useful.

The key takeaway: the dashboard rarely gets the arithmetic wrong. However, all too often it might be read as a single story when, in reality, it is holding four or five different stories, and the one that matters is usually the quietest.

AI ends here.

My message that isn’t in the AI piece, is, if designing a dashboard, first know what is its purpose. If you don’t know what it might reveal, check regularly for trends. As an example: when I first started counting headteacher vacancies in the 1980s, I didn’t know that church schools will be more likely to re-advertise their vacancy than non-faith schools. When looking at teacher vacancies; I didn’t know the rhythm for temporary and maternity leave vacancies being advertised, but I had a hunch, and added the field to the database. As a result, I would also say, collect as much data as you can, when constructing a database, as you might not know what it will tell you.

So, thanks to AI for simulating this post, and quickly offering me some thoughts about many posts I had already written. Does this make this a genuine post or an artificial one?

The ITT Census versus September acceptances

Posted on by
Reply

In my previous post, I said at the end of the post that  I would look at what any differences between the August ITT acceptance data issued by the DfE earlier today, and the ITT census to be issued in December 2026 might be, by looking at the difference between those accepted in September 2025 and the ITT Census of December 2025.

In reality, this task has proved something more  of a challenge that I had expected. This is mostly because of the late announcement of the targets for September 2026. These were not announced until April 2026, by which time many subjects had already made offers in excess of the new, and lower, targets. The fact that the targets for 2026 were in many subjects drastically reduced has made the analysis almost impossible.

Because of the reduction in the targets, it is possible that the outcome of the 2026 ITT Census could be the best year in terms of hitting targets since 2012, especially once Teach First and any other non-DfE portal recruitment is factored into the census.

However, there are a few subjects that rely almost entirely upon the DfE portal for recruitment that we now know are very unlikely to meet even the revised target for September 2026.

These are: Business Studies; Classics, Drama; Religious Education and the catch-all ‘Other’ that includes media and communication studies, social studies, citizenship and psychology.

Although how the ‘other’ subjects grouping reaches a target of 2,035 places remains a bit of a mystery. As I have advocated elsewhere, it is now time to breakdown the ’other category’ into its component subject targets to allow for meaningful discussions about teacher supply. With around one in eight places in the ‘other’ category and the category recruiting very poorly it is time for a more meaningful discussion about the classifications used by the DfE for its IIT targets.

A reminder of how the targets have changed this year when compared to the previous year

2025/262026/27
Total26,92020,800
PrimaryTotal7,650 5,520
SecondaryTotal19,270 15,280
Art and Design680605
Biology985675
Business Studies9001,200
Chemistry730690
Classics6075
Computing895565
Design and Technology965620
Drama620370
English1,9501,980
Geography935685
History790520
Mathematics2,3002,000
Modern Foreign Languages1,4601,085
Music565260
Others2,5202,035
Physical Education725655
Physics1,410810
Religious Education780450

 Some years ago, the total was nearer 40,000 trainees, so this has been a dramatic reduction.

When were secondary trainee targets last as low as 15,280? Not for some time. I will do  a search and see if I can find out the last time the targets were this low for both primary and secondary trainees. As these targets now include almost all routes, it will be interesting to see whether when they were last this low were there no ‘off target’ numbers such as the former employment-based routes?  

AI – a reminder from 2018

Posted on by
Reply

Sometimes it is worth repeating an earlier post. This one about AI and education was from as long ago as 2018. The post dealt with a Report by a House of Lords Committee. The most prescient section was; 

“One witness warned the Committee, “that the idealised world represented on social media “leads to many illnesses including eating disorders … and serious mental illnesses”.   The implication being that schools must put in place strategies to prevent such outcomes among future generations exposed to the perils of the modern world.”

You can read the full post at: AI and education – The view of the House of Lords Committee | John Howson

So, way back before Covid, Westminster was being warned about serious mental illnesses- that might impact schools, my highlighting.

Well, post covid, the SEND demand has not been generated by physical needs. The government seems to have largely ignored the warnings and is now trying to put ‘the stopper back in the bottle’, to use a phrase popular with the BBC this week.

Can we save on defence sending if we use some of the education budget to help the next generation work with the developments created by AI?

As Ukraine and The Gulf have shown, the mental capabilities of the armed forces may be more important than the physical attributes in a defensive war. We won’t be fighting imperial wars again – or I hope not – Iraq should have been the last one, if Afghanistan is seen as punishment campaign not a land grab, but without an endgame strategy that the Americans seem so bad at understanding is an essential part of any military operation or war.

I am not sure that the curriculum Review – remember that damp squib – really took on board the changes AI would bring to society. The Microprocessor revolution that changed the 1980s; the internet revolution of the late 1990s and the Jave Enabled smart phone revolution of the second decade of our century each changed children’s lives. AI has already done so, and will continue to do so in ways someone of my age cannot even imagine.

I have found turning a blog post into a webinar in just minutes awesome. Creating a poster from the same text is improving all the time, and its outcomes can now be regularly seen on LinkedIn and other social media platforms.

We are a tech savvy nation; but not yet a tech savvy education system. Is it time for bodies like The Royal Society to make their voice known, along with those that represent our creative industries, equally affected by AI.

I wrote in another post that ’the market porter of the early 20th century was replaced by the forklift truck driver. They in turn were replaced by the software engineer creating automated warehouses. Now AI will make most software engineers redundant, in the same way as the forklift truck driver has started to disappear as a role in warehousing.

Radical change is necessary in our education system. Do we need a national conversation around what that change should be, and not just a professional curriculum review?

Is AI bigger than the internet revolution?

Posted on by
Reply

During my lifetime, I have experienced three major revolutions driven by technology; the microprocessor revolution of the 1970s and 1980s; the internet and social communication revolution of the 1990s and 2000s around the internet and phones and gaming, and now the Artificial Intelligence revolution: or AI as it almost universally known.

Over the past week, I have set three different AI sites three different tasks, all using free versions of the software. The tasks were: draft a script for a play; create a website of this blog and turn a poem into a song and add new verses.

I was amazed at the results for all three tasks, and especially how quickly they were performed so early into the AI revolution.

I asked for a script for a play about a closure of a college, and ‘The last assembly’ gave me 5 key characters, four themes for the play and even some possible dialogue for a final speech and how the other characters reacted to its delivery.

In its way, even more impressive was the website created to market this blog. In under 5 minutes, and after a bit of interplay with the AI agent there was a useable website that those, I have sent it to thought was an attractive website. Sure, there were a few wrinkles to iron out, such as the wrong date being assigned to some posts used as examples, and the first version being just in monochrome. But, 5 minutes, and free, and perfection is expected!

The third task took a poem I had written to celebrate a music festival’s 40th anniversary and created a song for primary school-children based upon a first two verses from the poem that I had created.

Here was the poem’s first three verses.

Sing for the sound of singing

Sing for the joy of singing

Sing for the harmony of singing:

So, sing we all together.

As we sing together here,

Different schools, but single purpose,

Combining in this special place

For forty summers since.

Forty years of many pupils,

staff and parents, come all to

Celebrate in song and sounds

Old, new, would be traditional.

And my attempt at a song, something I have never tried to do before.

[C] Sing, sing, sing

[F] Sing for the joy of singing

[C] Sing, sing, sing,

[G] Sing for the sound of singing

  • So we
  • sing to·[C] gether.

[Am] Let us sing as one

[F] Different schools,

[C] but single purpose,

  • Let us
  • sing to·[C] gether.

Here’s what AI added

[C] Raise, raise, raise

[F] Raise up your voice in chorus

[C] Loud, proud, strong,

[G] Let every heart sing with us

  • So we
  • sing to·[C] gether.

[Am] Every voice belongs

[F] Every note a story

[C] Blending dreams

[G] in shared harmony

  • Let us
  • sing to·[C] gether.

[C] Hope, hope, hope

[F] Hope in the songs we’re sharing

[C] Shine, shine, shine,

[G] Shine with the love we’re bearing

  • As we
  • sing to·[C] gether.

I was also offered platforms where the backing track could be created. Now, the poem was written in an hour after receiving the invite, and isn’t great literature, so the song isn’t a work of art, but AI was asked for something primary school children might sing, and I think that’s what was created.

If I can do these three things with no tutorial support in less than half an hour all told, then we have to take AI, and its implications for our school system seriously now.

These three tasks were relatively value-free, but AI has the power to drive thinking, values and morals.

With the government extending the franchise to those aged 16 or above, what we teach in school is now of vital importance, and it must no longer be just a diet of facts or an attempt to create a purpose for handwriting other than as an art form. Politicians of all parties need to think seriously and quickly about what we need to teach in schools.

This blog was created by a human except for the verses of the song that were created by AI

AI in education: tackling the third revolution

Posted on by
Reply

Earlier this week, I sat in on a webinar hosted by the Education Policy Institute about Workforce Sustainability in the modern school system. The recording of the webinar can be accessed at Workforce sustainability in the modern school system Inevitably, much of the discussion was around how AI might make a difference to schools. AI is the third wave of the IT revolution, after the initial microprocessor revolution of the late 1970s and early 1980s, and the arrival of the web and the development of means to access it from desktops to mobile phones, and even watches, from the mid-1990s onward. AI has the possibility to significantly impact on the school system as we know it even more than the previous two ‘revolutions.

The impact might be in four areas

Recruitment – this will be cheaper, faster and more complex as both candidates and recruiters strive to make use of AI to help them secure either the perfect job or the best candidate, whether it be a teacher or any other post in a school.

Administration – compared to the days of pen and paper, typewriters and adding machines, technology in the past 50 years have vastly changed how processes are handled, and especially how data can be analysed. EPI even have a Model, described in the webinar, for assessing how MATs are operating. AI offers much more power to create systems with less need to burden teachers. Tracking individual learning will be far enhanced, well beyond what is possible even today.

Learning – AI could transform how students learn. No longer will teachers need to worry about coping with a range of abilities in the same learning setting: AI will tailor the learning package to the individual and make the learning experience stimulating enough to motivate every child. In doing so it could fundamentally change the role of a teacher, removing some of the drudgery and enhancing the personal interactions with learners. But, will the lightbulb moments teacher value so much disappear?

The contract between The State and families – will the current 190-day requirement to attend a ‘school’ or face sanctions that has existed since 1870 in England be replaced by a different sort of model where learning is at the pace of learner, and qualifications are obtained when ready. Could AI be used to identify those children not making progress, and offer support that families would be required to accept? I don’t expect such a radical change anytime soon, but it would be worth looking at how different groups in society see schooling today and what they want from it as the twenty first century enters the second quarter of the century.

What is certain is that the State needs to participate, and not leave everything to the market. There is a lot of profit to be made from AI, and schools represent a large potential market. The first step will probably be to agree on standards and certification for learning materials so individual schools and teachers can be sure what they are using are high quality learning materials. This is much more important that the debate over banning mobile phones in schools, but receives much less attention from politicians.

Big range in candidate’s chance of becoming a teacher

Posted on by
2

The latest ITT data for applications and offers for course starting this autumn was published by the DfE this morning. Initial teacher training application statistics for courses starting in the 2025 to 2026 academic year – Apply for teacher training – GOV.UK Normally, the data appears on the last Monday of the month, so this moth’s data is a week early. However, at this stage of the year the early publication probably doesn’t make much of a difference to any analysis of the data.

The good news is that the DfE allocations or targets, call them what you will, have already been exceeded in some subjects; even before any Teach First data has been added to these numbers. However, what matters is how many of the ‘offers’ turn into trainees on the ground when courses start. Based on previous years, it seems likely that this will turn out to a good year for the government and schools, but challenges still exist in some subjects.

At least four subjects, business studies; design and technology, music, and religious education won’t meet their targets this year. This is despite music having the highest conversion rate of applicants to offers of some 63%. By comparison, business studies has the lowest conversion rate of just 27%. It is possible that both classics and drama might also fall short of their targets this year, but the jury is still out.

SubjectTarget 2025/26June offersJune applicationsoffer to candidates’ ratio
Business Studies90025593427%
Chemistry730730243230%
Physics1,4101431476330%
Others2,520399117934%
Mathematics2,3002321621637%
Religious Education780397102939%
Biology9851275330339%
Total Secondary19,270161003848142%
English1,9501648393742%
Classics60419543%
Computing895895201544%
Art & Design680817165050%
Geography935854171250%
Design & Technology965587117450%
Modern Languages1,4601418272352%
Drama62026950154%
History790924166356%
Physical Education7251514263957%
Primary7,65087901527358%
Music56532551663%

For some subjects the ‘offers’ are well in excess of the targets/allocations, with physical education have offers double the requirement, and a high ratio for offers to candidates. The cash this will bring to universities through fees will no doubt be welcome, but is it a good use of taxpayers’ money? If the Teacher Supply modelling is correct, many of these trainees might struggle to find a teaching post in 2026: not a phrase I have written recently.

I am curious as to where the more than 4,700 physics candidates have come from? If those with offers turn up, then that will be really good news, but I think we need clarity about the numbers and their reliability in predicting trainee starters this autumn.

Although all regions have seen an increase in candidate numbers, the Midlands have seen a fall in candidate numbers as measured by the region of the training provider. There is an imbalance between provider regions with just three regions; London, the North West and the South East accounting for the bulk of ‘offers’. As some providers are located in one region but provide wider, and even national coverage, this should not be an issue, but is worth monitoring, especially in the subject that won’t reach their targets, for any regional shortages.

Still, for many admissions tutors in ITT this will be their easiest summer for more than a decade. 2026 might be even better.

Technology and Education

Posted on by
Reply

A recent event I attended, ahead of BETT 2020, led me to think about the place of technology in education. A simple typology would be to look at teaching, learning and support as three different areas where technology can be involved in schools. Of course, the first two are an arbitrary distinction, and technology can cover both teaching and learning at the same time.

I was interested to see the use of the term AI by many exhibitors at the HMC deputy heads conference I addressed last Friday about teacher supply matters. After all, TeachVac uses sophisticated and proprietary AI to handle job vacancies much more efficiently that say the DfE vacancy site that requires schools to upload every vacancy they have created.

AI is still at an early stage, and as a phrase can raise false hope of a new era for learning that are generally not yet justified. However, one only has to think of the rise of ‘contactless’ in the payment field to see the speed of change.

Contactless, as with smart phones and especially their cameras, demonstrates the problems of technology and inter-generational use. How many heads use contactless payments; how many teachers above 40, and how many teachers below the age of 40 don’t? The same can be asked about any photos taken during the summer break, and also how they were swapped; displayed or generally archived.

The speed of change has an important relationship to the power structure in schools. Are head teachers aware of what is happening and what represent good investment for the future and are they prepared to delegate downwards to those that understand technology and can make the case?

My first job as a teacher involved responsibility for hard technology in the school – 16mm and slide projectors, plus reel to reel tape recorders – and I recall asking for a video tape recorder to help both the drama and PE departments with their work. The first time the kit was used, the 6th form group entering the local one act paly festivals swept the board. They were a great group, but I hope seeing their rehearsals played back made them even better.

In the 1990s, I wrote that we were on the cusp of a revolution as profound as the introduction of printing in the 15th century. Looking back to the changes in the past quarter century, and how it has affected power relationships across the globe, I don’t think I was wrong. You only have to compare what is going on with Extinction Rebellion now with the CND protest of the 1950s and 1960s.

Scrapping BECTA in the Tory bonfire of the Quangos was probably the right thing to do; not replacing it with an advisory body on technology and education was a serious mistake.

I suspect that unless this blog post attracts attention, technology and the role of big tech and start-ups in education won’t feature in the general election: it should do so. Will 5 days a week and 40 weeks a year be the norm for schools for another 10 years, let alone for another 150 years it has been the framework for learning in this country?