AI@Autism & Classrooms of Future

Capacity Building for Nextgen Inclusion

The Centre for Policy Research and Governance (CPRG), a Delhi-based public policy think-tank and also a partner organization of the AI Action summit organized the two-day industry, government and academia event, 'पढ़AI: Conclave on AI in Education'. The conclave brand ‘पढ़AI’, an intelligent juxtaposition with India’s national language means study and was held on May 27-28, 2025 in Delhi. It was purposed for facilitating innovative AI-driven solutions that enhance pedagogy and promote an adaptive, inclusive, and future-ready education system.

Accessibility was a major theme and the session, 'AI and Inclusive learning' deliberated on this, of which I was a speaker.

Excerpts from My Talk:

A learning environment and inclusion in the context of Autism means Special Education, a tailored approach of needs and support that allows learners to achieve a level of academic success. It involves a plethora of supports and interventions in communication, sensory accommodations, visual and hands-on instructional strategies, structured teaching combined with Therapeutic portfolios.

Autism Data

Unfortunately for far too long, for the autistic population has been troubled by the legacy of low performance and infantile contents, with a near total absence of formal education. This needs to change and change exponentially, dramatically. The statistical significance of prevalence is huge.

The US has seen a 312 pc increase in the last 25 years, with 1 in 36 children being diagnosed autistic. This is the data from March 2024. What do the figures in India look like? An AIIMS research by Dr. Shefali Gulati studied a representative population sample through a method of population proportion by size sampling, so it was Kangra for hilly region, Palwal for rural, Dhenkanal for tribal, Hyderabad for urban and Goa for coastal. The study revealed a finding of 1 in 89. And this is the data from 2011 and if global trends are extrapolated, it can be surmised that the incidence would have risen in India also.

This is a very big population of affected children and we simply cannot allow them to fall through the cracks of the education system. Mainstreaming of autistic learners through formal education processes of subject learning, exam writing, certifications, embedding the values of conceptual clarity, digital proficiency and tool usage, working memory applications must happen. The old shibboleths of underservice have to be swept away.

What does AI do for Special Educators?

Special Educators spend a lot of time handling administrative workloads. When a child comes after an autism diagnosis, the teacher has to ascertain the present level of academic achievement and functional performance, create a descriptive of the child’s skill and ability levels, current challenges faced, and appropriate supports the student may need. Thereafter, the task of writing an Individualized Education Plan (IEP) begins. This must include specific, measurable, achievable, relevant and time-bound (SMART) goals and objectives for students. Preparing all this plus attendant student support plans are all time consuming tasks. It is here that AI becomes an indispensable paperwork partner. Not only does it reduce administrative workload, AI, data analytics and machine learning are most useful for adaptive learning environments since they are agile, responsive and dynamic. Personalizing learning experiences, which is at the heart of IEPs is taken to a different level in delivering differentiated content all at the same time reducing process latency and remarkably improving turnaround schedules.

With AI usage teachers can make lessons more accessible by leveling the text and create enormous amounts of practice modules. This refines assessment and decision-making processes, promising more accurate evaluations and insights into student progress.

For the teachers, the time saving accruals result in several high value gains. First, it makes teaching more effective by allowing teachers to focus on direct instructional engagements. Second, human centric conduct and direct student interactions improves. Third, teachers can customize content for maximum impact and student progress. And fourth, teachers have spare time to develop their own pedagogical skills.

Integration of AI into educational curricula is a necessity in SEN. Teaching about AI, equips students who struggle with handwriting and expressive language. This is suitable since most autistics do well with keyword approach. This method works well with LLM and VLM models.

Thus, tech notably AI catapults inclusion to nextgen by expanding the fundamentals such as personalized and self paced learning, problem- solving, working memory, interpersonal skills, lifelong and student centred learning. AI strengthens the SEN stack in the key area of critical thinking.

Besides teaching, AI helps teachers with advise on environmental management. SEN supports are at the heart of Inclusion. There are SEN support AI tools, which can inform teachers and support staff with general information on how best to support students with a specific learning need. It can be used to provide a general understanding of the needs/behaviors/adjustments or how the classroom can be more special needs friendly.

In all departments, that the SEN teachers are expected to perform - teaching coordination, review, designing and counseling roles – AI enhances the teacher, it is an extra pair of eyes and ears in the classroom.

Key Tools & Usage

Generative AI such as ChatGPT, Synthesia, Dall-E2 and Bard generate documents, create images, explain complex topics and provide step-by-step guidance thereby simplifying content for autistic learners. TeachMateAI, Magic School Generator tools improve task efficiency, personalize educational resources and help craft learning goals in a measurable way thereby promoting inclusion. Apps like Canva, Speech Blubs, Gemini Systems, are used by SLPs for speech and language developments, teaching social communication and conversational skills. Avaz, a popular India-based Alternative and Augmentative Communication app is beta testing its AI Assistant ‘Just in Time.’ This will automate for the user the process of folder creation and vocabulary selection. Once available commercially, Just in Time will instantly create topic folders in any language with just a tap using the power of LLMs.

So some of the skills, principles, formulas and processes deemed essential in classrooms today, including basics such as writing, grammar, calculations have no relevance as limiting factors. AI Assistants can make the practice of scribes redundant. Once usage is taught, apart from being an assistive technology, AI has huge EF spin-offs for the autistic individual.

For the Teachers, these tools saves time while prepping personalized lesson plans and sharpens focus areas - from TLMs Generator to Concept Explainers from any subject area appropriate to any pupil needs. And for older students, these can be farmed for developing Individualized Transition Plans (ITPs). The therapeutic uses of AI for social therapy have been known for sometime now. Social robots like NAO that imitate human forms of speech have been used in UK schools to help children with high levels of insularity; in the US, Milo – a humanoid robot that exhibits a wide range of facial expressions – is used to teach children to identify human emotions, a common form of Autism therapy.

Autistic children have been known to show to display high levels of comfort with computers; many even shun handwriting and leapfrog to touch-based Qwerty keyboards. Computer programmes are predictable, logical, and can provide an intellectual outlet. By developing relationships with AI and social robots, the hope is that ASD children will be able to better integrate socially with other people. The London Knowledge Lab has already tested this to some extent.

Designing For Future

This calls for the industry teaming up with special educators, therapists to create an extended ecosystem for classrooms, environmental transitions, group classes, social behaviors and skill improvements.

We need the IN devices to continuously capture critical data about displayed social behaviors and make quick, informal assessments to gauge if a child understands the activity/lesson plan or is making slight skill improvements.

Analyzing this data can lead to pinpointing social success factors and assessing a student’s strengths and weaknesses. Educators would be better able to understand the child’s learning and emotional development, slowly introducing them to more complex and varied social environments over time.

Accommodations & Attitudinal Changes Required

Classroom work and cognitive processes need to embrace and integrate tech. Multi stakeholder approach is needed. The Chief Learning Officers and Chief Academic officers in AI companies need to confer with Schools on curricula regarding Literacy development, Maths, Science, Leisure, Life skills teaching and Cyber security. Learner Variability Projects are required in schools.

The teaching courses of B.Ed and M.Ed must be both - about AI and Teaching with AI. They should include modules on AI literacy, principles, concepts, applications, limitations, usage precautions, ethical implications and advanced technology modules. There can be a specialization of B.Ed (AI) also. The World Economic Forum’s Future of Jobs Report 2023 predicts a growing demand for education-related roles, including vocational education teachers and special education teachers.

What SEND and Specialist Schools need to do? First, promote AI literacy by integrating it into curricula. Second, create an AI Lead team which will see oversee implementations in therapeutic and academic divisions. Third, build capacity by supporting educators and staff in integrating AI into teaching, learning, school management and operations by sending them for professional development programmes. Fourth, to advance AI in education pedagogy, funds for device ecosystem and AI tools have to be made available. Fifth, provide guidance by establishing usage guidelines and ensure ethical principles of student protection and data privacy.

Few countries are proactive in harvesting the enormous possibilities and we have examples of AI policy interventions for schools. Countries are framing basic principles on how to integrate AI and digital literacy in classrooms. In the UK, the Office for AI is currently conducting research to support primary and secondary schools to teach critical skills such as the limitations, reliability and potential bias of generative AI; how information on the Internet is organized and ranked. Australia, Japan and New Zealand have also outlined guidance around teaching with and about AI. This is applicable for all student categories.

Closing the digital skills gap

For identifying gaps, tech and AI skills gap audit is required. Ensuring equitable access to these sophisticated tools is paramount. The cost of these technologies, coupled with the necessary infrastructure like high-speed Internet and training costs could create AI haves and have-nots. AI works best for teachers who know both – SEN and AI. A teacher has to be really structured about doing it and for this successful case studies are needed. Integration of human expertise, AI capabilities and comprehensive public-private partnerships are needed for planning, implementation and impact assessment.

The Education Minister at the AI Conclave

The AI conclave, which had five knowledge sessions, was addressed by the Education minister Dharmendra Pradhan. In his valedictory speech, Mr Pradhan said, ‘integrating AI in school education is no longer optional, it is essential.’ Mr Pradhan said the transition from chalkboards to chipsets is imminent in a big way and spoke of various government initiatives – Centre of Excellence in AI, leveraging AI in Indian languages to promote critical thinking in classrooms. For policy recommendations on AI, he asked academicians and tech experts to get their act together.

The minister has been evangelical in egging on apex educational institutions for tech and AI adoption. He has to do this continually for there exists big pockets of reluctance, doggedness to cling to legacy methods, a low aptitude for proactive adoption, irrational fears about teaching creativity being dampened and tech misuse by students. The red flags of concerns, as Mr Pradhan said, can be handled by our country’s human intelligence that will lead the AI revolution.

Postscript

For Inclusion, AI needs to be seen as an indispensable accessibility tool for bringing about key transformations in primary and secondary education to promote better student outcomes. At the heart of all this however is the pivotal premise: Educating the Educators! CPRG, with its mission of ‘AI for Global South’, please note.