What impact might Generative AI upon the classroom? A publisher’s perspective

Leave a reply

Introduction

Many people have written on this subject, but I thought I would bring a perspective from the providers of support to teachers and learners.  There are three lenses that I think will be useful in considering this problem:

  1. What have we learned from the experiences of Covid, and the need for human interaction to support the wellbeing of learners
  2. How should education shift in an information rich world?
  3. What have we learned from the past impact of new “transformative” technologies? 

Human in the loop

It is without doubt that generative AI has emerged as a transformative force in K12 education and beyond, offering unprecedented opportunities for personalized learning and creative exploration.  However, we have also just emerged from the largest experiment in digital enabled learning in human history, and we are beginning to learn the lessons from home learning and Covid.  Primarily, across the world and in developed and emerging societies there is crisis in pupil wellbeing.  Attendance is down, scores in international comparative tests are stalling in many countries and support for children’s mental health is under unprecedented pressure.  Alongside this, many schools are rolling back some of the technology platforms that supported learning through lockdowns in favour of face to face learning.  What does this tell us?  Face to face teaching has a significant impact on the wellbeing and attainment of pupils.

Therefore we need to ensure that technology, and AI in particular, are supporting the “blended” approach to delivering learning, with the teacher and classroom at the heart of learning. 

Moving from a knowledge rich to a skills based curriculum

We have established that face to face delivery is a core element of the future delivery of learning.  But what is the learning that is being delivered?  Commentators and educators have long argued that current model of delivering learning is based on a 20th century model.  There is little value in the current focus on teaching and assessing declarative knowledge, when many questions can be answered by asking Alexa, and more complex questions are being increasingly well answered by AI.  Rather it is increasingly important that we focus on the skills required to function in society – much has been written on this, but core to the AI discussion are the skills of deconstructing a question into a prompt, interpreting the output of AI and synthesising a new and personal response from the output.  Alongside the “traditional” 21st century skills, (collaboration, communication, creativity, critical thinking etc) we need to build a curriculum that enables learners to contribute effectively to society.

How long will this take?

As I wrote in my previous blog post it is useful to compare the noise around generative AI to the two other technologies that were predicted to transform education in the last decade- adaptive learning and blockchain.  The hype around Knewton, and the vast amount of investment raised, did not replace the teacher as many predicted.  Adaptive learning platforms are now a part of many courseware solutions, but are no longer the focus of the investment community.  Similarly, the crypto/ blockchain revolution was slated to change the way education and credentials were delivered.  This may happen, but there is no real sign of it yet.  However, I believe that AI, and Generative AI in particular will have an impact on learning.  It is useful to refer to the Gartner Hype Curve:

https://en.wikipedia.org/wiki/Gartner_hype_cycle

Currently we are still attaining the peak of inflated expectations.  For an education technology to be successful it is not enough for it to enhance the learning experience.  It needs to be embedded in the ways that learning systems work, support curricula and enhance the assessment systems that passports learners through their lives.  Many technology stumble on this – they may be effective, but they cannot be blended into existing classroom approaches.

And also note there are a vast array of ways that AI will support education – replacing routine tasks, lesson planning, drafting reports and home school communications, and supporting whole school reporting.  These and other uses should massively increase the efficiency of educational institutions.

In conclusion

Generative AI has the potential to revolutionize the way we approach education. By leveraging adaptive algorithms, it can create customized learning materials that cater to the unique needs and learning styles of each student. This technology can generate practice problems, interactive stories, and even simulate scientific experiments, providing students with a rich, engaging, and diverse learning experience.  However the human teacher’s role is irreplaceable. Teachers bring empathy, understanding, and a personal touch that AI cannot replicate. They are skilled at interpreting students’ emotions, providing encouragement, and fostering a supportive classroom environment.  And the curriculum and resources hneed to transform to prepare learners for a world where they will harness the power of AI.

Notes

I am merely scratching the surface of how AI will transform education here – I am sure I will write more on the subject.  But I am indebted to many sources of information and insight in this area – including:

Also, I asked ChatGPT 4 via Windows 11 CoPilot for its thoughts on the subject.  Some were useful and are reflected in the above.  However, as will be the case for some time the blog post has been unrecognisably transformed by the addition of my experience, thoughts and narrative style.  I hope.

Edtech in 2024

Leave a reply

Thoughts on the last decade in Edtech

Over the next few months I am re-entering the world of consulting – more on that elsewhere, but I am excited by the prospect of helping clients in the edtech and wider technology space navigate the increasingly noisy world of digital technology.  You will see that it is a decade since I last posted my thoughts on this blog – a decade where I have been engaged in supporting schools digitally around the world, a decade where technologies have come and gone, and a decade where the impact of Covid is still playing out.

I thought I would take the opportunity to restart my blog by sharing some thoughts on the state of the EdTech sector as we enter 2024.  I will look at three key areas – the impact of Covid, the impact of shifts in the technology landscape and how the narrative is being shaped by PE/ VC investments.

Firstly, as has been widely noted, the Covid 19 pandemic has driven a shift in the take up of technology in schools.  For years, education has lagged other industries in using technology to deliver its core product, driving improved learning outcomes (Note I will expand on this is a future blog – schools are highly technology enabled in other areas).  However, Covid has driven a step change in digital take up:

Originally published in The Adaptation Advantage: Let Go, Learn Fast and Thrive in the Future of Work, by Heather E. McGowan and Chris Shipley, with a Foreword by Thomas L. Friedman.

These changes fall into three areas:

  1. Creating a digital environment for schools.  The overwhelming majority of schools now have a digital environment within which the manage both the school infrastructure and communication and collaboration with learners.  And the vast majority have plumped for Microsoft of Google environments.  All suppliers to schools need to work out how to integrate with these systems
  2. Workflow enhancements – schools are now using technology in many more areas of the school ecosystem, – from managing staff, to improving communication with stakeholders and the provision of online parents evenings
  3. Delivering digital learning and monitoring attainment.  Throughout Covid learning went online with a variety of levels of success – some tools delivered well (e.g. simulations, online “tests” and video content), others were clearly a stop gap (ebooks/ worksheets).  But the majority of schools are now able to support learners at home to some extent.  The education system is more resilient

However, as we move out of the pandemic, towards a “new normal”, some of these changes are more embedded.  (1) and (2) above are here to stay whilst there is a shift back to blended learning.  More schools are using flipped learning to maximise the value of face to face time, and many are looking to digital formative assessment to support differentiation in the classroom – but these are based around the classroom model. In the absence of conclusive evidence from the sector, this is based on schools’ experience of what is working. 

Running alongside this narrative of a changing classroom is the current hysteria around generative AI.  Again, I will look at this in more detail in a future post, but I believe there are three areas where Generative AI will have an impact on suppliers and educators:

 SuppliersEducators
Workflow enhancementsWill be able to use AI to enhance customer support, better market portfolios and to improve education services – e.g. in delivering marking servicesSchools will be able to use AI to better understand the wider needs of their learners, and communicate more effectively with stakeholders
Content generationWill be able to build content to support customers, and build quality content on the fly from reliable content banks.  Suppliers will also be in a position to quality assure content generated in this wayTeachers will be able to build learning content around specific requirements – however it will need quality assurance
Transforming learningUnderstanding learners needs based on rich datasets owned by education suppliers has the potential to support teachers and learnersWorking with suppliers and internal datasets educators will be able to provide more effective learning to pupils

It is important to note that I believe that traditional content owners will continue to have a major role to play.  Whilst schools can generate content, publishers, exam boards and edtech players will be able to “kite-mark“ learning content and provide coherent programmes.

I compare the noise around generative AI to the two other technologies that were predicted to transform education in the last decade- adaptive learning and blockchain.  The hype around Knewton, and the vast amount of investment raised, did not replace the teacher as many predicted.  Adaptive learning platforms are now a part of many courseware solutions, but are no longer the focus of the investment community.  Similarly, the crypto/ blockchain revolution was slated to change the way education and credentials were delivered.  This may happen, but there is no real sign of it yet.

Which brings me to my last point – the market for provision of digital solutions that truly support learning is being distorted by PE/ VC investments.  Driven by technology trends and by the need for relatively short realisations of return, the investment community seeks start up players that aim to use these technologies to disrupt education.  And schools’ expectations are heightened by the hype around these services.  However successful companies (traditional and startup) know that it is a long game with tight margins, that trust is at the heart of education supply, and that switching resources in and out is hard.  And moreover, that success in the education world still is driven by old fashioned skills such as supporting teachers in managing classrooms, building a great channel to market and providing content that educators can trust.

New Primary Assessment Framework

Leave a reply

As many will you have noted, DfE published their response to the consultation on primary school assessment yesterday (27th March 2014) – this can be found here https://www.gov.uk/government/consultations/new-national-curriculum-primary-assessment-and-accountability.  This post summarises those changes, and briefly considers their commercial implications.

As widely predicted, decile reporting has gone, however, there have been several changes to the assessment regime:

  • New baseline assessment at Reception.  This is to be sourced from a range of assessment approaches (including commercial) approved by DfE.  This will be optional – schools not choosing to make this assessment will be assessed on progress from KS1 in 2022, and only on attainment beyond (see floor standards below)
  • KS1 from 2016 new externally produced and internally marked tests for maths and reading will inform teacher assessment, and a SPAG (this last to be externally marked?) will inform TA of writing.  These tests will return a scaled score (around 100).  New performance descriptors to aid TA will be published in Autumn 2014
  • KS2 New performance descriptors published.  Tests as now, but producing a scaled score.  Children’s results reported to children alongside school, local and national averages
  • Low attaining pupils – no change to p-scales

New floor standards introduced:

  • Schools judged on whether schools make sufficient progress from KS1 (2022) or from the baseline if taken (2023 on) OR
  • 85% or more of pupils meet expected standard in reading, writing AND maths (to be similar to 4b)
  • Explicitly stated that schools not meeting floor standards ”could result in the school becoming a sponsored academy”

League tables will publish more data:

  • Average progress made by pupils in reading, writing and mathematics;
  • Percentage of pupils achieving the expected standard in reading, writing and mathematics at the end of key stage 2;
  • Average score of pupils in their end of key stage 2 assessments; and,
  • Percentage of pupils who achieve a high score in all areas at the end of key stage 2.

What does all the above mean for publishers/ those seeking to support schools?  The obvious main implication is that there will be a surge in demand for assessments in Reception.  These need to be available in short order, as it is likely that schools will begin to trial approaches in September 2014.

However, there are wider implications.  Schools will need to track both progress and attainment across Reception, KS1 and KS2 – as all do now.  Unfortunately, this response does not detail how progress will be measured, and merely notes that scaled scores will be available at KS1 and KS2, with detailed progress descriptors.  These progress descriptors will need to be mapped to approaches enabling schools to determine whether children are below, on or ahead of track.  And tracking software needs to be adapted to record and analyse this data.  One of the key things Ofsted will be looking for in the future is a robust approach to this, and schools will be looking for validated approaches.

Thus the opportunity is twofold – both in terms of developing resources and the systems to enable schools to ensure that they are on track to exceed the new floor standards.  And publishers have little time to meet them …

Is the new focus for education data not content?

Leave a reply

It has been apparent for some time that there may be a structural decline in the content market in schools. However, there is one area that has remained buoyant over recent years – data.  This is demonstrated by the recent market interest in the acquisition of assessment provider Granada Learning – an acquisition that could have been seen as defensive by a publisher, but is seen as an opportunity by Investcorp.  Why is this?

Data is increasingly essential to the operation of schools.  Both assessment/ attainment data and management data are generated in ever increasing quantities, and both have the potential to drive school improvement.  It is used for accountability purposed through Ofsted; top communicate with parents and most importantly teachers use data to enhance the learning they deliver to pupils, for example in the assessment for learning process:

Data is essential to this cycle – curriculum data describing what should be taught; pupil data collecting data about the child and the school’s interaction with it and attainment data gathered from tests and teacher assessment.

Few schools make effective use of this data – it is often manually entered into spreadsheets and shared at particular points in the school year, and many teachers are not equipped to analyse it.  Therefore supporting school data use is a significant area of opportunity for the commercial sector – in terms of:

  • Data acquisition – supporting schools in generating attainment and other data through assessment and from content tasks
  • Data management – helping schools effectively to collect and share data
  • Data analysis – Supporting schools in making the data useful to teachers in terms of supporting pupils, and identifying successes

Publishers have a significant opportunity to enhance their portfolios by using their content to generate useful data; and by developing systems to consolidate and analyse data.

Grade inflation – is it the fault of publishers?

Leave a reply

Another week, another tumultuous period for education policy … with significant implications for the structure of educational publishing.  Hot on the heels of the release of the first draft of the primary programmes of study, Michael Gove’s desire to re-instate some backbone to 14-16 education by returning to “O” levels was mysteriously leaked to the Daily Mail.  Ignoring the issues that this alone might raise (see this very interesting FT blog for one perspective), the ideas when fleshed out advocate a single exam board for each core subject and a stripping back of the National Curriculum (Gove’s desire to remove it completely is impractical as it would require primary legislation) – all this to be achieved by 2014.  Interestingly, this “leak” came the week before the scheduled release of the Select Committee on Education published its report on 15-19 examinations.  This specifically looks at the issue of competition in the exam system, and the place of educational publishers within it.

The key issue that Gove, the Select Committee and many recent curriculum reviews and media commentary is trying to address is that of grade inflation – and attendant confidence in the exam system.  And a key issue that has been identified driving this inflation is competition between exam boards.  The argument is that in order to attract schools, exam boards make exams “easier” and provide substantial support to students to protect their market share.  Proponents of a single board per subject approach argue that this would remove this competitive pressure.  The Select Committee considered this, and was concerned that a single board would stifle innovation.  It looked at three functions of exam boards that are differentiators:

  • Syllabus – what an exam covers
  • Exam administration – how the exam is delivered, marked and analysed
  • Support – the support that the exam board gives to schools and students

The Select Committee argues that “Syllabus” is not necessary for competitive advancement of the exam system, and could be developed by “National Subject Committees” regulated by Ofqual (any similarities to the old QCA are purely accidental).  Competitive pressures in Exam Admin and Support could (with effective regulation) continuously enhance the exam system.  The report then goes on to examine each of these areas in detail.  And of particular interest to publishers is the section on Support.

There are two areas of support considered.  First is traning – seminars given by exam boards to schools and pupils.  This is an area of concern after the recent Telegraph expose, and the report commends Ofqual’s decision to end the practice (which opens up a potential opportunity for publishers).  However the section on textbooks is of greater interest to publishers.  There is discussion of the links between textbook endorsement, the curriculum covered by the textbook and the drive to raise pupils’ grades.  The argument is that exam board endorsement of a particular text book (whether produced by an external company or a partner company in the case of Edexel and Pearson) stymies competition in the development of resources in that subject.  In turn, the need to gain competitive advantage by gaining endorsement narrows the content of that textbook to closely follow the desires of the exam board – which as noted earlier is driven by a requirement to drive up grades to retain market share.  Additional factors such as authorship by examiners (are they the best authors?) and the close links between certain exam boards and particular publishers also caused the Committee concern as they may have issues for the quality of resources.

Interestingly, and perhaps astutely, the Education Committee chooses not to lay the blame for these perceived quality issues purely at the door of the relationship between endorsement, examiner authorship and competition between exam boards,  In fact the main driver is identified as the school accountability system – where schools are being assessed by their GCSE grades their is an intolerable pressure on them to raise them – and this inevitably leads to a focus on raising grades, and getting pupils over the D/C boundary.  It recommends that “school system” accountability be separated from the exam system – and measured through PISA like sampling – this should take the sting out of the “falling standards” argument.  And in order to address the narrowing of teaching within schools, they should not be assessed on exam grades alone.

A key question is – how accurate is this analysis? In a competitive market resources are created to meet market demand.  And in this case the key driver is the concern uppermost in customers’ – schools’ – minds, that of achieving their targets.  Thus necessarily resources are created that match the syllabus which is being taught, and if “endorsement” and “examiner authorship” increase this perception these are the legitimate tools of the publisher.  And in any such environment “good” resources and “poor” resources will be developed.  It is up to schools to choose those which best meet their needs. The unfortunate truth is that this is not a “pure” market as it is skewed by government policy.  The current accountability system for schools ensures that for all players in the system the overriding driver is meeting exam based targets.  And this creates an environment where the “best” resources do not necessarily win.  The question is – will the current education reform proposals reduce this focus on exam grades?  In their current form it appears unlikely.

 

Curriculum change again

Leave a reply

Last week saw the publishing of the draft programmes of study for Michael Gove’s revised primary curriculum.  Coming on a back of a series of reviews (the Rose Review and the Cambridge Primary Review) at the end of the Labour Government, the publication is as ever surrounded by controversy.  But what does it mean for publishers?

Curriculum change offers publishers an opportunity – new curriculum structures require new teaching schemes and resources – and demonstrate the complex relationship between Government, practitioners and publishers.  Without the commercial sector, Government education policy would not be delivered so easily, as publishers translate curriculum change into resources that support teachers in delivering that curriculum.

So what are the key changes in the new curriculum?  There are several headline changes:

  • The detail of what is to be delivered remains prescriptive.  Whilst not radically different, has an increased emphasis on knowledge and rote learning
  • The programmes of study are now yearly rather than in two year blocks
  • Levels and level descriptors are to be abolished (however details of the replacement assessment regimen have yet to be published)

Publishers can adapt to changes in curriculum content easily.  It is the second two changes to teaching practice that will be harder to manage … the change to annual programmes adds a level of prescription – not only are teachers told what children should know, but when they should know it – this leaves less flexibility for children to catch up.  And the removal of levels is potentially the most disruptive … for some time now the focus on assessment has been criticised from all sides.  However, levels are not just used to tell others where children are, but to inform teaching in the classroom.  Teachers use them to track individual progress and understand when interventions are required – and increasingly with the adoption of assessment for learning techniques levels are the language with which children understand their own targets, progress and assess themselves and their peers.

Techniques such as APP and Assessment for Learning have supported the teaching profession in understanding how to assess where a child is, and how to support them to make better progress.  They have become a shared language for the profession and changing this will present a significant challenge to teachers, and the publishers supporting them.

Thus the latest round of curriculum change does present an opportunity to publishers – but also a threat.  Until we understand how children are to be assessed, and how teacher will use this to inform teaching it is unclear what form alternative teaching resources will take, and prudent publishers will watch this space keenly.

Is making digital content the easy bit?

Leave a reply

There are many great digital products out there.  However, there are many more potentially better products that have been created that have not made it commercially.   This is the perennial Betamax versus VHS debate – products that are not necessarily better win out because of a variety of factors – and it seems that digital publishing is no different.

I have personally been involved in the development of many digital products.  Some have been commercially successful, but have not been innovative or provided clear advantages to their users over traditional media, whilst others which have had the potential to make significant changes to the way a customer segment operates have struggled.  The objective “quality” of a product does not necessarily correlate with its commercial success.  Why is this?

Firstly the value proposition of the product needs to be clear and simple.  The more a product does, the harder it is to communicate to customers why they should be interested.  This is why many simple, derivative products are successful.  They may be “dull” – but customers get them.  But even a product with a strong value proposition can fail.

Potentially more importantly there needs to be a way of communicating this value proposition to the market.  This is made up of a combination of strong sales and marketing input throughout the product design and development process, and ensuring that your sales channel, or channel partners, are capable of selling the product – they may have excelled in traditional media but may not have the expertise to deal with the new product.  And finally, customer expectation must be considered – is this what they expect from you?  And if not, are you in a position to re-educate them with the risks that this may hold for your brand?

In many ways developing the product and the value proposition are the easy bit – they remain entirely within the control of the publisher.  Once a product is released into the market it becomes the to some extent the property of customers – and they are not in your control.  Sales and marketing is the increasingly sophisticated but still blunt weapon at your disposal to influence them.

 

Can we be trusted with data?

Leave a reply

A report into the usefulness of league tables across the public sector has just been published by the British Academy.  In it Harvey Goldstein and Ben Folen argue that tables do not well serve the purposes put forward for such data, which are:

  1. ‘public accountability’- enabling the funders of institutions to determine whether they are receiving value for money
  2. “supporting choice” – helping the users of services to chose a school or hospital based on its track record
  3. “control” – to provide levers for government to influence the activity in institutions
The problem with league tables as they stand is that not only do they create perverse side effects (the prime purpose of a primary school seems  now to be to ensure it is well placed in the league tables rather than giving a well rounded curriculum), but the very indicators used to define league tables are limited by what can be measured – and it is very hard to measure qualitative things.
More importantly data in league tables is published as fact, with none of the caveats or reliability information around it supplied.  Such data is merely a sample.  When schools look at assessment data they are usually given the confidence interval around it – the range of scores that a particular result could be indicative of.  Such data is not published around league tables, and due to the small sample of a single assessment the confidence interval (or range of possible results indicated) is large.
The Government is seeking to make league tables in education more useful by publishing more data.  However this does not make them more reliable.  It also does not address the issue of data literacy in the target market – parents and teachers.  How many actually understand what Contextual Value Added (CVA) actually means?  And how many just look at the position of a school in the league table and take it as a firm indication of the quality of all aspects of a school?
But this issue of data  literacy is evident across the profession.  It is often considered that the more data that is supplied to schools the better decisions that a school will make.  And the data provided in systems such as RaiseOnline is very detailed.  However there are very few teachers who are able to take advantage of this data.  Across education there needs to be a focus on making data rich and usable, rather than publishing it because it is there.
Harvey Goldstein, one of the authors of the report, was interviewed on the Today Programme on March 29th 2012 – the segment is at 1h32mins here – catch it while you can!

The demise of Encyclopaedia Britannica – is this the inevitable end for all print?

Leave a reply

It has been widely reported today that Encyclopaedia Britannica is ceasing publication of its print version after 244 years.  In the face of competition from initially Microsoft’s Encarta (remember that?) and now Wikipedia, the investment of thousands of pounds in a static repository of knowledge for individuals and institutions no longer makes sense.  But, in a world of Kindles and Apple’s iBook Textbooks, does this point to the inevitable demise of all things print?

The short answer is, I believe, no.  Firstly there is the historical perspective.  As has been noted many times over the years no one new media “class” has replaced another.  Television did not replace radio.  Digital audio has yet to supplant analogue (in fact vinyl sales have been growing recently).  And the internet has not replaced linear TV.  What seems to happen is that each new media makes the mix of consumption more complex.  (Note that within genre of media, there can be replacement – DVD replaced video, and will eventually be replaced by digital streaming – but video delivery remains – ditto the replacement of illuminated manuscript with print …).

However, within print certain classes of “misfit” publishing will vanish.  Encyclopaedias are one such class.  For what is an encyclopaedia other than a database in a book?  It has been crying out for the invention of digital to add the functionality (e.g. cross referencing) that it has tried to implement in a ham-fisted fashion on paper … this was clear from the early days of electronic legal publishing where even traditional, hide-bound barristers found themselves buying databases such as CELEX and the All England Law Reports.

Where does that leave other publishing formats?  Well, an ebook does not add much functionality to a novel – other than portability at the expense of fragility.  My expectation is that the two will co-exist for generations more.  The case against the traditional textbook is easier to make.  These are repositories of knowledge.  They are multimedia – containing images (and would be enhanced by video), contain exercises and assessments, and so are prime for a gradual decline.  However, until the hardware delivering an electronic textbook becomes cheap, reliable and ubiquitous (and Apples iBook for iPad is none of these yet), this cannot happen.  But the market is aware of the potential disruption, and this is causing a systemic decline in sales.

So is print on its way out – no.  But are there more classes of print publishing that will go?  Without a doubt – and the textbook is likely to be one of them.

Gove seeks to sever the links between exam boards and publishers

Leave a reply

Ofqual have launched their consultation into education resource development … this will have an impact on the way that education publishers and exam boards conduct themselves…

Over recent years one of the ways that publishers have gained competitive advantage for their resources has been to gain exam board endorsement.  And this has driven relationships such as Pearson’s with Edexcel.  However, in the light of many representations to Goverment and the recent telegraph expose of exam board practice, Michael Gove is looking to sever this relationship.

This will have an impact both on the way that publishers construct their courses (perhaps making them more board independent) and how they market them.

For details of the consultation, please see the Ofqual site:  Qualification support material and services: call for evidence.