Connecting curriculum, assessment and trust – qualifications for the next decade
A speech delivered by Sir Ian Bauckham, Chief Regulator at Ofqual, at the Confederation of School Trusts annual conference.
Good afternoon. It’s a privilege to be able to speak to you today.
My job as Chief Regulator of Qualifications is to safeguard the national asset that is our qualifications system.
Why do I say: A national asset?
England’s qualifications enjoy a global reputation, and are a central part of the soft power that this country has internationally.
GCSEs and A Levels also have a place in our collective consciousness at home: today’s teenagers’ parents did GCSEs, and their parents and grandparents may have done A Levels.
But qualifications also play a critical role in shaping the educational experience students have, and so much of the work we do as educators (and I speak as one).
And they have played a critical part, along with other key features of our system, in achieving England’s strong educational performance that we have seen recently, for example, in PISA.
But you don’t need me to tell you that we live in a changing and often challenging world.
Maintaining a world-class qualifications system does not mean standing still.
With the future in mind, I want to touch briefly on 3 topics.
V Levels. Assessment on screen. AI.
Let’s turn to V Levels first.
The introduction of V Levels from 2027 will create a stronger vocational route for post-16 students.
They will sit alongside A levels and T Levels.
A. V. T.
This will be the new level 3 post-16 landscape.
So who or what are V Levels actually for?
V Levels will be level 3 qualifications, so in the same ‘tier’ as A Levels.
V Levels are not like T Levels, in that they are not full-time courses, preparing students for entering a single occupational area. Which is what T Levels are.
Actually, only a minority of teenagers know with certainty what they want to focus on later.
Most want, to some degree at least, to keep their options open.
A V Level is a smaller qualification, to be taken with 2 other V Levels or A Levels to make a tailored study programme.
It introduces the student to a vocational field and the range of roles that exist in that field.
Unlike the T Level, it does not require an industrial placement.
V Levels will be entirely teachable in school sixth forms as well as colleges.
And, they could be one possible progression route for students who may have taken the recently announced vocational GCSEs in key stage 4.
Now, I am often asked: isn’t a V Level just a rebranded BTEC?
Well, let’s look at what a BTEC actually is. BTEC is a private proprietary brand name owned by Pearson. Only Pearson-owned courses can be called BTEC.
Pearson determine what the content of a BTEC is.
I, for one, don’t think it is right for critical decisions about what a large proportion of our 16 to 18-year-olds learn to sit in the hands of a private company.
It’s not a principle we accept at GCSE: all GCSEs, regardless of which exam board offers them, are required to have the same content, determined by our democratically elected government.
I don’t see why post-16 students taking vocational qualifications shouldn’t get the same guarantee.
And while BTEC is the best known of the brands currently, it’s not the only one. Other boards offer level 3 vocational courses too.
But here’s a thing: you can’t directly compare the grades from one to another.
We would not accept this for GCSEs or A Levels.
I don’t see why it is acceptable for post-16 vocational students either.
It is these types of longstanding issues that the V Level reform will set right.
And I for one think that is a good thing. In fact, in the world of vocational qualifications, this is a revolutionary reform.
So, the first 3 V Levels will be in education and early years, accounting and finance, and digital systems and data, and will be available for first teaching in September 2027.
The second, much larger tranche, will follow the next year.
Now, looking to the future is not just about what qualifications are available.
It is also about the how.
And the how must include how qualifications work in the fast evolving world of technology, including AI.
So let’s start with one apparently straightforward question: should exams move on screen?
It’s no secret that some exam boards would like us to “be braver”, “move faster”, to prevent us “getting behind the curve”. They say so regularly.
But my job is not to do what exam boards want me to do.
Qualifications are a public asset. So the public has to want and trust innovation of this sort.
And it turns out that simple assertions that say “the world is going digital so it’s about time exams caught up” pose as many questions as they answer.
We do sometimes have a tendency in this country to assume we must be somehow ‘behind the curve’ internationally.
“We should look at what is happening in other countries” is often the appeal of those who think that.
OK. It’s true that New Zealand, Israel and Finland have gone further than most.
All very small systems, and, crucially, all with long term investment plans for IT in schools.
But other respected and well-resourced systems have paused or reversed. Norway, Sweden and Denmark, for example.
Actually it turns out that there are very few examples of whole large systems of high stakes qualifications (and ours is a large system) moving on screen.
Sometimes it’s concerns about actual deliverability: insufficient confidence it will work on the day.
But, increasingly, it’s wider concerns about ‘screens in schools’, or ‘screens in childhood’.
We must also consider fairness and social justice. There is differential access to, and therefore familiarity with, different kinds of devices and platforms. The digital divide.
I also need to think about the readiness of schools to introduce large scale computer based exams.
And, dare I say it, the capacity of exam boards actually to run large systems reliably enough to be trusted with the aspirations of hundreds of thousands of teenagers, and their nervous energy on the day.
Not all exam board ventures with large tech systems end happily.
There are of course also some potential advantages to onscreen assessment.
Some tasks may be more easily presented on screen than on paper, and some special educational needs might more easily catered for – though even that is a far from uniform picture.
We are looking at all these issues as we decide how we will regulate the arrival of on screen assessment in England’s schools. Expect to hear more in due course.
And now to the question of AI.
In relation to qualifications, there are 3 main areas for us to think about here.
Firstly, will we still need qualifications that prove what students know and can do when – we are told – AI will be able to do everything anyway?
If the answer is no, then game over. Not just for qualifications but for education too.
But if the answer is, as I will argue, yes, then the second question is:
What should we do about the misuse of AI in qualifications; AI-driven cheating?
And thirdly, to what extent should we let AI run the exams and qualifications system for us? And what would the implications of that be?
So let’s take those 3 questions in turn.
Do we still need qualifications that assess and prove what people know and can do?
Of course, the workplace of the future will be AI enabled. It is fast becoming so already.
But that cannot be allowed to mean no more educated humans are needed.
The future workforce will need higher-level expertise to augment, or add value, to work done by AI.
Specialised knowledge, finely balanced judgement, will be central to that.
If you can’t do something additional or different to AI, then by definition you will not be able to add value, and your work prospects will be limited.
So we can rapidly put to bed the argument that we can just give up on teaching, learning and assessment. Education in other words.
Colleagues, we won’t need less education, we will need more, and we will still need qualifications both to frame it and accredit it.
So to the second question.
What should we do about the misuse of AI in qualifications?
AI cheating in other words?
This is simple. We cannot allow it.
Extended writing coursework is vulnerable to cheating, as I have said many times before.
And every time I say it, I am told to look at the ‘marvellous’ things happening in universities. How many have gone beyond ‘old-fashioned exams’ to ‘embrace’ ways of assessment more ‘suitable’ for the AI age.
Well, I beg to differ.
Just recently I noted that Princeton University, a prestigious Ivy League university in the US, has moved to invigilate all in-person exams, amid concerns about AI and academic integrity.
On this side of the pond, UCL Law School last year announced an overhaul to make assessments more “AI‑proof” by making greater use of in-person invigilated exams.
Indeed, universities are finally waking up to the harsh reality that if they want their qualifications to continue to be taken seriously, then assessments do need to tell the truth about their students’ actual achievements.
For qualifications to be worth the paper they are written on, we cannot allow them to be subverted by an assumption that ‘AI slop’ adequately replaces the fruit of genuine human endeavour.
And we cannot allow this to wash back into schools and colleges either.
While AI – properly used – is helpful to ease the burden on teachers, and to improve the quality of curriculum and teaching resources, outsourcing student thinking to AI does not develop the human mind.
It leads to less learning, less satisfaction, less usefulness in the workplace, and ultimately is an assault on the realisation of human potential.
It is less cognitive outsourcing – it is more like cognitive surrender.
Research is showing this again and again.
Where students use AI to shortcut learning, they achieve less well than those who go through the effortful process of thinking and learning for themselves.
You don’t get fit by watching a marathon on TV.
Over the past 2 decades we have sleepwalked into allowing young people undiscriminating access to smart phones. Many spend hours on addictive social media and communications. Just look at the harm that has wrought.
We need to learn from that and not repeat the mistake with undiscriminating use of AI.
My third question was: should we let AI run our qualifications system for us?
There are 2 ways of approaching this question.
The first is an argument from accuracy, efficiency and cost.
The second an argument from ethics.
There are certainly areas where AI can improve efficiency, cut costs and increase accuracy.
Helping to write exam questions, quality assurance, analysing performance, and so on.
But we often also hear arguments, mostly based on claims about accuracy, for AI taking over marking.
To be clear, I am talking about AI marking here, not machine marking of questions with a very limited range of easily defined right answers.
AI marking, by contrast, is where we ask AI to mimic human judgement of complex responses. Essays, artworks, performances.
We need to be clear that using it in this way opens up real ‘routes to harm’.
AI hallucinates, and it reflects biases. All current research supports this.
But colleagues, even more importantly, generative AI, by definition, has the potential capacity to ‘improve’ itself. To act autonomously.
So it could extend a human-agreed mark scheme to deal with cases not covered by that mark scheme.
And at that point the expert human is no longer in the loop, even indirectly.
And we must confront an important moral question.
Qualifications – indeed education – exist for humans, for human benefit.
Not the other way around.
In my view the dignity of the human person requires that consequential judgements made about a person’s future need direct human oversight.
And – just imagine – if AI marking started to legislate for itself on ‘edge cases’, given the wash back effect of qualifications into teacher behaviour – we effectively set up a route for generative AI outside human control to influence decisions about what teachers teach students.
Moreover, using generative AI in marking in this way also makes the basic right to contestability very difficult, if not impossible, to exercise.
You may be frustrated when a review of marking means that an examiner error is uncovered and a grade has to be corrected.
But at least you can challenge human errors.
How do you challenge a judgement made by a self-evolving black box that no-one really understands?
So in short, I think the AI question goes beyond efficiency, accuracy and cost savings, important and desirable as those may be.
This is fundamentally an ethical question.
A question about the extent to which we want to outsource decisions about the transmission of our civilisation, and therefore the very future of our civilisation, to AI.
Of course AI has a place.
But we must remain in charge.
We must legislate and regulate to be clear about what AI can be allowed to do, to bring about genuine benefits for humans, and what it absolutely cannot.
Colleagues, let’s not forget what, and who, education is for. It is not to create an illusion of learning, to con young people into believing they have achieved.
It is to genuinely realise human potential, yes, of course, for the workplace, but first and foremost to enable young people to take their place in and contribute to our culture, our civilisation, which is what it means to be fully human.
It is imperative that we all, as educators, are vigilant and exercise the wisdom needed to safeguard that precious enterprise.
Thank you very much for your attention.