November 2, 2023

The Future of Assessment: How AI Grading is Transforming Corporate Learning, Apprenticeships, and Further Education

Manjinder Kainth, PhD

“It is increasingly obvious that we are on the cusp of a revolution in artificial intelligence that will be no less profound than the arrival of the printing press or the Internet,” ran the leader in the New Scientist in the “AI revolution” special issue, April 2023.

Of course, educators aren’t late to the races. They’ve been using AI for some time. “A third of teachers now use artificial intelligence (AI) tools such as ChatGPT to help with their work, double the number who used it five months ago,” reported Schools Week (14 September), dispelling the myth that educators are late-adopters and it’s the kids who are always the first off the mark.

ChatGPT: 1 in 3 teachers use AI to help with school work (

Teachers are using AI to create resources, plan lessons, write reports and risk assessments, respond to parents’ emails, compose UCAS references and job adverts, and more besides.

Well aware that large language models are not perfect, they’re even using that to their advantage. To produce model essays, they know they have to adapt what ChatGPT writes or, better still, ask students to have a go, opening up discussions about what makes for coherent, cogent academic argument.

As Professor Becky Allen, co-founder and Chief Analyst of the teacher survey tool Teacher Tapp, recently told ResearchED, “I like to think of them [large language models (LLMs) like ChatGPT] as being like the best kids we teach in our class. They get the A*s, they get over 90 per cent. But they don’t get everything right.” Ultimately, however, she is excited about “how transformative” they can be, helping educators “in reducing their workloads and helping them make better plans and better resources”.

And what about assessment?

Let’s unpick how AI grading will impact corporate training programmes, apprenticeships, and further education.

Saving precious time and money

AI grading has huge potential to ease workload and improve the lives of educators.  

Graide, for instance, observes and learns how each question is marked, and then is able to replicate it at scale, ensuring greater speed and efficiency. Teachers no longer have to mark the same answer again and again, ad nauseam.

AI also enhances the quality of grading.

In fact, the median grading times were reduced by 74%, and the number of words of feedback given increased by a factor of 7.2, when we compared grading on paper to the Graide platform. We also estimated that a university, with 3500 STEM students, using Graide could save over £240,000 a year.

With 20+ students in college classes, speeding up the whole process matters in further education and corporate training or apprenticeships.  


University of Birmingham - Graide Efficacy Research

Immediacy is of the essence

Large groups sizes slow down assessment and feedback when timing, or rather, immediacy is of the essence.

The Education Endowment Foundation guidance report on “Teacher feedback to improve pupil learning” made 6 recommendations, the second being “Deliver appropriately timed feedback that focuses on moving learning forward”.

Dylan Wiliam, Emeritus Professor of Educational Assessment, Institute of Education, UCL, wrote in his guest forward for the report:

“For feedback to be effective we need to create classrooms where students welcome and use feedback. As Stiggins, Arter, Chappuis and Chappuis remind us, the most important decisions taken in classrooms are not taken by teachers but rather by learners.”

Timely feedback is imperative for supporting learning, motivating students, and ensuring student engagement. Timely advice and action plans allow learners to put things right straightaway, rather than forget an assessment or, worse, worry about their result.

Supercharging the quality (and quantity) of feedback

Once the AI observes an educator’s response to a particular type of answer, it can replicate it at scale, increasing the accuracy, quality, and quantity of feedback.

That is immensely exciting.

Having large class sizes can disincentivise writing detailed comments for each individual when it takes up so much time.

Graide’s Replay Grading not only means having to mark one type of answer just the once – it also replicates feedback, placing it centre stage in the learning process.  

Alternative angles to take, model answers and study tips all go a long way to address students’ difficulties or stretch the most able.

Joining the dots – spotting patterns

Its power to deal with enormous volumes of data, to spot patterns and to analyse errors and misunderstandings in a particular knowledge space, that’s where AI comes into its own, exceeding what humans can ever hope to calculate themselves.

Using embedded knowledge graphs alongside generative models means you could take a question, give it a sample answer and then classify the different techniques that have been used within answering that type of question.

This is where it gets really interesting – and hugely beneficial to educators.

You are then able to generate questions which tackle similar problems.

In the past, people have used automation tools to create randomized questions by changing variables. But randomization has its problems: you can change the difficulty level significantly quite by accident.

But the beauty of embedding knowledge graphs into these solutions is that we can create bespoke learning pathways for each individual, outlining next steps to take, pitching future tasks to their exact ability, and tailoring the next assessment to address weaknesses.  

That’s an enormously powerful and exciting thing, a dream that I’ve had for some time now.

Harnessing AI enables formative assessment at scale, placing it at the heart of the learning process, allowing students to practise, practise, practise at a level that suits them.

And large language models like ChatGPT will only get better and better at creating these formative assessments from educators’ seed questions.

Final thoughts

I guess we ain’t seen nothing yet. If you think how far we have come from rudimentary text complete on email to current iterations of LLMs using billions of parameters, imagine the transformation down the track.

Exciting times, indeed, and it’s great to see educators and academics leading the charge.

For more on the future of AI in education watch our round table, recorded in March this year:

AI & Higher Education - Is it time to re-think teaching and assessment? - YouTube

Or download the full report:


Further reading

Benefits of using a Digital Grading Platform to Evaluate Student Learning (

 Further watching

TE Summit - Using AI to reduce grading workload and increase student feedback - Manjinder Kainth - YouTube

AI & Higher Education - Is it time to re-think teaching and assessment? - YouTube

Manjinder Kainth, PhD
LinkedIn logo.
I’ve been intricately involved in teaching for many years. I have over 6 years of private tutoring experience, taught in higher education, and most recently worked on designing and delivering crash courses for a high school.