THIS IS THE SECOND IN A SERIES OF THREE ARTICLES ABOUT HOW AI WILL IMPACT SPECIAL NEEDS TEACHING IN THE NEXT FEW YEARS AND HOW EDUCATORS CAN ENGAGE WITH IT TO MAKE THEIR LIVES EASIER AND MORE PRODUCTIVE.
A Path To More Productive Teachers and Better Outcomes For Pupils
Small steps
There will be lots of big shiny AI-based 'solutions' coming out of tech company workshops over the next five years. The 'try this cool new app' selling approach has worked for 30 years and will work in most industries (and maybe, to some extent, in mainstream education), but I think that apps focusing on the SEND sector might need to take a different approach.
AI is not, strictly speaking a ‘tool’. It is an ‘agent’ able to make decisions on its own. This makes it scary for most non-techies. And we can include most special needs teachers in this category.
So, the way to get special schools up the AI learning curve effectively will be by introducing AI generated additional features into their existing software application, step by step, over several years, so that each step make sense and achieves ready acceptance from educators and leads them down a path towards fuller AI implementation - as far as they, and their leaders, are comfortable going.
It is possible, in this way, to envisage a point in the not too distant future when AI based systems carry 50%+ of the full teacher workload for almost all special needs pupils.
Let’s be specific
In the previous post we listed a whole range of general teacher tasks that AI could assist with - or, ultimately, just do.
Here, we are going to be more specific and list So, what specific tasks will these AI based systems help teachers and school leaders with?
Let’s start by looking at these.
Planning
EHCP outcomes
- Suggesting the next year’s EHCP required Outcomes for each individual pupil, based on historic performance data and comparative data with other pupils around the country with similar disabilities as well as historic individual performance data provided by the school and other information provided by clinical, parental and social support.
Provision mapping
- Identifying cohorts for whom previously utilised interventions might help to improve academic and specialist outcomes. Matching resources with interventions.
Target setting
- Devising an evolving set of smaller annual and termly targets for each pupil, based on the agreed EHCP Outcomes and on the pupil’s previous achievements and inclinations.
Lesson plans
- Building a similarly evolving sequence of individual lesson plans for each term, based on this information and on the success or otherwise of previous lesson plans.
Delivering
Lesson content
- Providing the actual individualised lesson content for each lesson, from millions of sources worldwide to maximise engagement and performance – to be delivered by the teacher or online.
Measuring performance in each lesson
- Based on each pupil’s interaction with the lesson content provided. Offering a mark for each session which informs the grading decision. Adapting for each pupil, from day to day.
Grading pupils
- Suggesting any grade changes when measuring performance against a curriculum framework or target adjustments. Highlighting grades which teacher might like to review?
Teaching evidence
Faster, more comprehensive, record commentary
- Proposing commentary on evidence records, based on images, video or work recorded by teachers. The suggestions would be based on a million other records made by teachers teaching similar stuff to similar pupils - and on teacher corrections to previous records.
Moderating
- Select and edit specific evidence records to show teachers 'What this level looks like' at other schools. Saves those evening moderating meetings. Developing teacher training based on these.
Keeping track
Tracking
- Illustrating pupil performance against teacher expectations and/or the performance of similar pupils elsewhere.
Feeding back
- Feeding back to teachers to help improve lesson design and delivery.
Reporting
Conventional reports
- Providing suggested commentary for yearly, termly and other reports.
- Reporting on individual performance to school leadership, parents and other relevant interested parties, using multi-media.
EHCP annual reviews
- Producing all the required reporting and other documentation required by each local authority, in the form required by each one. When combined with suggested new Outcomes, this will make annual review preparation much easier and the meetings more productive.
Parents
- Monthly summaries for parents/carers of their child’s performance and achievements - aiming to encourage parents who are not ordinarily engaged.
- Suggestions to parents about what can be done at home to reinforce teaching at school.
Analytics
Group comparisons
- Comparing class, year, group performance against other similar groups, previous years, other schools or anonymised standards and producing customised vivid, useful school-wide analytics, effortlessly.
- Monitoring individual and group performance against EHCP targets set.
Highlighting performance
- Picking out good or poor performance. Suggesting improvements based on this.
Leadership interaction
- Answering specific questions from school leaders or governors.
Caring
Attendance
- Recording and reporting. Relating attendance to performance.
Behaviour
- Recording and reporting. Suggesting remedial action. Following up. Summarising.
Safeguarding
- Informing. Suggesting remedial action. Tracking follow-up. Summarising.
The calculation of each of these useful outputs would be informed, not only by the individual performance, preferences, responses and actions of the pupil in question, but by relating these to:
- the experience derived from similar data across the world
- from the measured performance of thousands of other children with similar conditions and similar targets within the software
- a constant stream of inputs and corrections applied to its draft suggestions, from those using the system, including teachers, parents, carers and other professionals.
Because the whole process is based on 'machine learning' each new element introduced will likely be clunky at first, but will refine its decision making every day, until it becomes a vital aid to achieving higher pupil/teacher ratios and reducing staff stress levels, while still improving standards.
Will this mean a big reduction in the need for special needs teachers?
No
There will still be plenty of need for specialist staff of the right type and calibre. But if this, in its full flowering, were to mean that individual special needs teachers could deal with 50% more pupils, the potential cost saving to the exchequer, in the UK alone, would exceed £3 billion a year, out of a total annual budget of £11 billion.