‘AI-powered, teacher-facing tools are much needed innovations’

AI-driven tools could be a real game-changer in the remote assessment process, write Dr Parves Khan, global research and insight director at Pearson; and William Cronk, Pearson’s digital content specialist

Teacher workload is a growing crisis

Pre-COVID-19, teacher workloads were already reaching breaking point. In 2019, the Ofsted Teacher Workload survey found that a quarter of teachers in England work more than 60 hours per week during term time – 12 hours above legal limits set by the European Union. A National Education Union (NEU) report in 2018 warned that two in five (40%) teachers are planning to quit the profession by 2024 due to unbearable workloads and poor work-life balance.

Now, with all the catch-up necessitated by the pandemic, excessive workload risks becoming a national crisis unless a solution is found. So, the question is – what will this solution look like?

At Pearson, we’re trialling studies with edtech partners to test the role AI could play here. As the capacity and variety of AI expert systems increases, and the natural language processing capabilities of AI expands, new possibilities for AI in education will continue to emerge. One exciting growing area is AI-powered auto essay-marking; tools like this are desperately needed innovations to tackle the mounting pressure of excessive marking.

Technology as an enabler

Artificial intelligence in educational software is not a new development, but the new capacities of AI are allowing for more meaningful applications. There are a number of commercially available AI-powered auto-marking applications already available – but the vast majority focus on objective and rigid subject matter like mathematics and computer sciences. These are well-suited for online formative assessment activities undertaken using closed questioning techniques, such as yes/no questions, multiple-choice questions or drag-and-drop activities. While valuable, this is a rather narrow way to do formative assessment.

“As the capacity and variety of AI expert systems increases, and the natural language processing capabilities of AI expands, new possibilities for AI in education will continue to emerge”

Our solution enables teachers to pose open-ended questions which can be automatically analysed and assessed by a computer. The ability to offer real-time feedback to teachers and students is another key feature we’re developing.

Our pilot: trial of new essay auto-marker

We worked with teachers from nine schools who provided close to 500 English Pearson Edexcel GCSE past paper essays for us to use in the trial. The majority of these were answering a 6-mark English language question. We worked with an edtech start-up, Progressay, who specialise in auto-marking long-form English essays. Progressay fed the 500 answers through their marking engine and we compared results to those from Pearson’s senior examiners. As well as providing a raw mark, the marking engine gave a detailed feedback report on each answer, explaining why the marks were given.

Key findings:

  • The teachers were happy with the marks awarded to their students – broadly in line with their expectations
  • They were impressed by the level of detail in the feedback they received with each mark – beyond what they were expecting
  • Surprisingly, most teachers in the trial weren’t too worried about the agreement rate for a number of reasons:
    • They understood the size and scope of the trial, and that the agreement rate will increase as we do larger trials;
    • They understand that teachers don’t always agree with each other in terms of marking, and that as long as the engine is explaining why the marks are given (intelligent correcting), it can be a really useful tool for everyday marking;
    • And finally, they understand that even if the accuracy doesn’t improve, the automated feedback can still be a useful tool for intervention planning and writing skills. 

No, the robots are not coming!

AI-powered, teacher-facing tools like auto essay-marking could be a game-changer in assessment – not just in terms of reducing the stress of marking workload, but in terms of improving the learning and teaching experience. By reducing manual marking, teachers will be freed up to provide more engaging interactive classroom lessons, to spend more time guiding and motivating students, to use real-time assessments and feedback to monitor student progress, and have much more time to explore specific strategies for struggling students.

Such tools will elevate, rather than diminish, the critical role of the teacher – they will have more time to do what they are trained to do: devote time and expertise to teaching young people.

Getting from the future potential to reality of tomorrow will take concerted time, effort, will and investment. The most encouraging aspect of this trial is the fact that all the teachers were enthusiastic about participating in future trials and helping to develop this product to make it as useful and accurate as possible. Our trial will continue, with a larger supply of answers and a broader range of data moving forward.


You might also like: Developing a growth mindset? Game-based learning could be key


 

Leave a Reply

Upcoming webinar

The digital student journey

How to include everyone from recruitment to graduation and beyond

Wednesday, May 19

11AM (BST)