BATOD/BAEA/NDCS – DSA assistive software consultation
BATOD, BAEA and NDCS submitted a joint response to the proposals for the way in which the government fund assistive software through Disabled Students’ Allowance (DSA).
“British Association of Teachers of Deaf Children and Young People (BATOD )is the only professional association for Qualified Teachers of Deaf Children and Young People (QToDs) in the UK. The Association represents the interests of QToDs and the children and young people they teach with a range of governments and other agencies. The Association supports QToDs and organises continued professional development (CPD) courses and national and regional meetings to provide relevant upto-date information and to disseminate good practice. Strong links are maintained between BATOD and the UK governments, as well as voluntary bodies, especially the National Deaf Children’s Society, in order to contribute to policy development in this field.
Contact Information: Teresa Quail [email protected] National Executive Officer and Qualified Teacher of Deaf Children and Young People
British Association of Educational Audiologists (BAEA); the only professional association in the UK for Educational Audiologists. Educational Audiologist are typically QToDs who have completed an additional post graduate qualification in Educational Audiology.
National Deaf Children’s Society (NDCS) is the leading charity for deaf children and young people dedicated to creating a world without barriers for every deaf child and young person. We give expert support on childhood deafness, raise awareness and campaign for deaf children’s rights, so they have the same opportunities as everyone else.
Contact Information: Martin McLean Deputy Director, Policy and Campaigns
BATOD, BAEA and NDCS appreciate DfE’s position that the setting has a responsibility to make reasonable adjustments to meet the needs of individuals they serve and the proposal that “if the software products freely available to the student do not meet the student’s disability-related support needs, and the support required is within the scope of DSA, paid-for software may still be funded for the student”. Some deaf students will require specialist software packages as they are reliant on a high level of accuracy in the live captioning content. The BATOD/Jisc Captioning Reference group for deaf learners has a ‘Live captioning in education’ guidance pending publication.
BATOD, BAEA and NDCS welcome the confirmation that “assistive technology training on all software (free or paid-for) agreed in a student’s needs assessment report would continue to be provided where required to students through DSA”. Currently the DSA applications for deaf students only involve clinic audiologists, often from adult services, who may not have significant experience of the educational needs of deaf students. BATOD, BAEA and NDCS would strongly recommend evidence from an Educational Audiologist or QToD is sought and accepted. These professions will have been involved with the young person for a number of years and will have contributed to the transition planning. Their role will have entailed providing the training for effective use of the assistive listening technology, acoustic environment and unbiased informed sharing to support the young person’s informed decision making.
This joint submission frames the majority of responses in the context of live captioning.
We wish to highlight the information from the Jisc’ Video captioning and accessibility regulations’ webpage “Public Sector Bodies Accessibility Regulations (PSBAR) require that audio and video recordings made available from 23 September 2020 must meet accessibility standards. These standards are primarily defined by the Web Content Accessibility Guidelines (WCAG) 2.2 Level AA.”
“The regulations in the UK apply to pre-recorded, not live, content. PSBAR allows for up to 14 days after the video has been published, for captions to be added. Although this is minimum guidance, and [setting] policy could be to add captions more quickly.
WCAG 2.2 AA does require live video to be captioned. However, when the PSBAR was introduced in the UK, this requirement for captioning live video was omitted.”
- Are the AI features that have been incorporated into DSA-funded assistive software products compliant with Higher Education Programs’ (HEP) policies on AI?
BATOD, BAEA and NDCS consider that it is not currently possible to determine sector-wide compliance because higher education providers (HEP) AI policies vary significantly between institutions and are changing rapidly. The BATOD/Jisc live captioning surveys also indicates that provision is inconsistent across further and higher education, with use often dependent on individual need, staff initiative, infrastructure and local policy. DSA decision-making should therefore avoid assuming a single, stable sector-wide position on AI. Compliance should be assessed in relation to the specific tool, function, task, data-processing context and disability-related access need, rather than by treating all AI-enabled tools as equivalent.
Reference information:
Live captioning in UK further and higher education: exploring the current landscape – Jisc website
Live captioning in education: the view from schools and specialists – Jisc website
- How should we take into account HEPs’ policies on AI in our decisions on what to fund through DSA?
BATOD, BAEA and NDCS recommend that HEP AI policies inform, but do not determine, DSA funding decisions. Such policies must be considered alongside Equality Act duties and the student’s assessed disability-related needs. Where an AI-enabled tool provides a reasonable adjustment, the appropriate response should be to agree a transparent, controlled and policy-compliant use case, rather than to refuse support by default.
The input of a specialist such as a QToD or Educational Audiologist should be sought where the assistive technology trainer does not have a specialism in deafness to support training in the use of AI related technology used to support the needs of deaf students. Preparation for Adulthood throughout education should include ongoing training in using AI related technology so students are familiar with its capabilities and how to use it in advance of starting Higher Education.
Reference information:
BATOD/NDCS Specialist Deaf Curriculum Framework – Preparation for adulthood
- Should DSA fund software products containing generative AI tools that can create original content for students’ academic work?
BATOD, BAEA and NDCS do not support a blanket exclusion of software products solely because they contain generative AI features. Funding decisions should distinguish between assistive functions that support access, comprehension, organisation, communication or stamina, and functions that generate substantive academic content. Any funded use must be clearly bounded, transparent and compatible with institutional academic integrity requirements.
- Are there any ethical concerns regarding the use of AI in assistive software products? Does this pose any risk to students?
BATOD, BAEA and NDCS identify significant ethical risks, including data protection, confidentiality, intellectual property, bias, accuracy, hallucination, academic integrity and unequal access. These risks are heightened where students upload lecture recordings, unpublished research, sensitive personal data or assessment-related material into tools not supported/approved by institution’s IT system. For deaf students, and for students who rely on captions, transcripts or summaries to access spoken delivery, inaccurate AI outputs are not merely inconvenient; they can create a direct access barrier and undermine equitable participation.
- In what ways can AI be used to improve support for disabled students?
BATOD, BAEA and NDCS recognise that AI can improve support where it reduces clearly identified access barriers, including live captioning, transcription, note-taking, text simplification, tone checking, planning, organisation and task management. AI can also support visualisation of ideas and organisation of notes and text according to personalised individual needs. For deaf students, accurate live captions that have a scroll-back functionality, and reliable transcripts can support access to teaching, learning, revision and independent study, particularly where listening and concentration fatigue, acoustic conditions or specialist subject vocabulary create additional barriers. However, AI must be implemented as part of a supported access strategy, with appropriate training, safeguarding, quality assurance and specialist oversight. Effective use of AI, particularly live captioning, requires the specialist input of specialist assistive technology (assistive listening technology and good quality microphones and devices), human note takers and BSL interpreters, QToD or Educational Audiologist input, training or individual needs assessment. Note: for higher education the QToD works in the capacity of a Specialist Support Professional.
Reference information:
Live captioning in education guidance (publication pending)
Assistive Listening Technology Working Group/NDCS – Quality standards for assistive listening technology
British Deaf Association (BDA) Discission paper on AI BSL
Live captioning in UK further and higher education: exploring the current landscape – Jisc website
Live captioning in education: the view from schools and specialists – Jisc website
Scottish Sensory Centre training
Scottish Sensory Centre (SSC) assistive technology hub – contact SSC hub lead Dr Brian Shannan
- In what way do AI tools designed for general use differ from those integrated into assistive software for students with disabilities, and how do any differences affect user experience, accessibility, and so on?
BATOD, BAEA and NDCS consider that general AI tools and AI integrated into assistive software should not be treated as interchangeable. General AI tools are designed for broad consumer use; AI within assistive software is more likely to sit within an accessibility workflow and be developed around disability-related barriers. This distinction matters for deaf students, for whom reliability, accuracy – particularly with accents, atypical speech production eg ‘deaf voice’, complex subject specific terminology and abbreviations, compatibility with assistive listening technology, usability, training pathways and accountability are central to effective access. Free or mainstream tools may be appropriate in some cases, but they can introduce cognitive load, compatibility issues, usage limits, data risks and inconsistent performance. Caution must be applied to the statistics used for accuracy as they will be for a typically hearing population. Some deaf students may have limited or no auditory discrimination and may rely on the visual live captioning. HE staff, needs assessors and budget decision makers should take into account the views of deaf students in the HE setting to make informed decisions.
In addition the co-occurring needs of some deaf students requires input from other guidance eg text modification and access to live captioning on a range of device options.
Suggested other services – UK Deaf Student’s Union, Jisc, RNIB, RBID, Patoss, British dyslexia association
- Do you have any comments on our proposal to fund demanding software only where there is no suitable non-demanding software available that meets the student’s needs? (This means software that needs a higher spec computer to run)
- Do you have any comments on the definition of “demanding software” for DSA purposes?
BATOD, BAEA and NDCS urge caution in adopting this proposal. Software should not be excluded from funding solely because it is technically demanding if it is the most effective way to meet an assessed disability-related need. The definition of “demanding software” should be transparent, evidence-based and regularly reviewed, and should reflect real educational use rather than minimum technical specifications alone. It should also avoid creating a route by which students lose access to necessary higher specification equipment because an apparently less demanding tool is judged to be adequate in principle but not in practice. This is particularly applicable to live captioning software products to be used by deaf students for the reasons discussed above.
- Do you have any comments on our proposed approach to composition and mind mapping software? (There are sufficient free versions available to they won’t be funded).
- Do you have any comments on our proposed approach to note-taking, recording and captioning software? (Free versions are available so won’t be funded except for exceptional circumstances).
BATOD, BAEA and NDCS do not support moving note-taking, recording and captioning software to an exceptional-circumstances model without a stronger evidence base.
Better understanding by the HEP workforce is required about the Equality Act and Public sector body regulations for subtitles on recorded media which does not relate to live captioning.
These note-taking and captioning tools are core access arrangements for many deaf students, including those who rely on accurate live captions, transcripts or structured notes to access spoken teaching and to review content independently. BATOD, BAEA and NDCS emphasise that notetaking and captioning tools might be additional to human support, informed by the needs of the individual. The Jisc live captioning survey indicates that live captioning use is growing in FE and HE, but provision remains variable and is often dependent on individual need, staff initiative and local infrastructure. It also identifies practical challenges around accuracy, infrastructure, staff awareness and policy gaps. These findings do not support an assumption that free or mainstream captioning tools provide reliable parity with specialist provision. Inaccurate, delayed or unavailable captions may directly restrict access to teaching, participation and assessment preparation. Free tools may also be time-limited, weak with accents or specialist terminology, inaccessible, unsuitable for long teaching sessions, or incompatible with institutional data and AI policies. Where software is removed, students may require more expensive and less independent human support.
It is not clear how an ‘exceptional circumstances’ approach would work in practice. It is in a deaf student’s interest that they have access to the best quality captioning available. There is no method of determining sufficient comprehension of captioned content other than the student’s self-assessment from their experience of using automatic captioning within an education environment. Universities are very different environments to schools and FE colleges. What might have been sufficient for a small online group may can be very different to what meets a student’s needs when attending a large lecture theatre or practical class in person.
Reference material
Live captioning in UK further and higher education: exploring the current landscape – Jisc website
Live captioning in education: the view from schools and specialists – Jisc website
BATOD/Jisc Reference group ‘Live captioning in education ‘guidance pending publication
- Do you have any comments on our proposed approach to OCR software? (will continue to be funded).
- Do you have any comments on our proposed approach to presentation support software? (not funded as already available)
- Do you have any comments on our proposed approach to research and referencing software? (Readily available to all students so won’t be funded).
- Do you have any comments on our proposed approach to revision software? (Readily available to all students so won’t be funded).
- Do you have any comments on our proposed approach to software to support vision impairment? (No changes to current funding)
BATOD, BAEA and NDCS support the proposal to maintain current funding for software to support vision impairment and recognises the co-occurring VI profile of some deaf students. This reflects the importance of specialist, reliable and accessible software that accounts for the full profile of needs rather than the suggested primary need/disability. However, the same needs-led rationale should also apply to students whose disabilities are less visible but whose access needs are no less substantial.
- Do you have any comments on our proposed approach to speech-to-text software? (we will only fund speech-to-text software for students who, as a result of their disability, rely on speech recognition as their primary method of written communication and of controlling a computer)
BATOD, BAEA and NDCS consider the proposed threshold to be too narrow. Some deaf students may have actively used ‘read out loud’ as part of their normal way of working through primary and secondary education stages to support the self-checking and editing of their work. As noted above, deaf students may have other co-occurring needs and may require speech-to-text because of pain, fatigue, coordination difficulties, fluctuating conditions, processing barriers or reduced stamina, even where speech recognition is not their sole or primary method at all times. Speech recognition-based provision is also affected by accuracy, infrastructure, acoustic conditions and implementation quality. Funding should therefore be based on functional impact, context and assessed need, not an artificially high dependency threshold.
- Do you have any comments on our proposed approach to text-to-speech software? (we will no longer fund text-to-speech software other than in exceptional circumstances for the following categories of student:
- students with dyslexia
- students with ADHD
- students with autism
- students with mental health conditions
- students who are Mac users (as Macs have advanced in-built text-to-speech software)
We will continue to fund text-to-speech software for students with a vision impairment.)
- Do you have any comments on our proposed approach to time and task management software? (Not funded as software is widely available)
- Do you have any comments on our proposed approach to training software? (no changes, still funded).
BATOD, BAEA and NDCS support the continuation of funding for training software. However, the consultation appears to assume that assistive technology training can fill gaps created by the removal of funded specialist software. This assumption needs to be tested. Training is only effective where the tools themselves are suitable, accessible, reliable and available for long-term use. Training should take into consideration the integration of specialist hardware and may need a bespoke approach delivered by a specialist with extensive knowledge of deafness rather than non-human training.
- Do you have any comments on our proposed approach to typing tutor software? (not funded as free to access are available).
- What types of exceptional circumstances do you think should be considered when deciding whether to make exceptions to the proposals set out above?
BATOD, BAEA and NDCS do not agree with the premise of exceptional circumstances. Paid for captioning software should be available until there is confidence among users and the supporting workforce that free software meets the majority of deaf student needs. Some free or mainstream tools may be inaccessible, unreliable, inaccurate, insecure, incompatible with other assistive technology, insufficient for long or specialist academic tasks, unsuitable because of fatigue or cognitive load, or prohibited by institutional policy. Access to paid for captioning software should explicitly include deaf students, and other students who require accurate captioning, transcription or structured note-taking to access spoken teaching. The Jisc live captioning survey points to accuracy, infrastructure, staff awareness and policy gaps as practical barriers. These barriers should be explicitly recognised within any exceptions framework. The framework should also recognise fluctuating conditions, co-occurring needs, communication differences and barriers to self-advocacy.
- What are your thoughts on our proposal to fund the lowest cost software product available that meets the needs of the student? (recommending type of software rather than brand as there is a range of prices on the market).
BATOD, BAEA and NDCS do not consider that lowest cost should be the primary determinant. The test should be whether the software meets the student’s assessed needs in practice, including accuracy, accessibility, usability, reliability, training, technical support, compatibility, data protection, continuity and performance in academic contexts. For deaf students, a lower-cost captioning or transcription tool that is less accurate, less reliable or weaker with specialist vocabulary may not meet the access need at all. Availability alone does not guarantee effective access; accuracy, infrastructure, staff awareness and policy conditions all affect whether a tool works for students. A lowest-cost approach risks false economy where unsuitable tools lead to non-use, appeals, reduced independence or greater reliance on human support.
- What do you think are the equality impacts of our proposals?
BATOD, BAEA and NDCS are concerned that the proposals risk negative equality impacts for disabled students, particularly those with non-visible, fluctuating, multiple or complex needs. For deaf students, and for students who rely on accurate captioning, transcription or note-taking support, reduced access to specialist provision may directly affect access to teaching, participation, assessment preparation and independent study. Jisc’s live captioning survey shows that use is growing but provision remains inconsistent, with practical barriers around accuracy, infrastructure, staff awareness and policy gaps. The proposals may therefore widen digital poverty, increase variation between HEPs, disadvantage students with regional accents or specialist academic vocabularies, and place additional burdens on students least able to navigate appeals or exceptional-circumstances processes. BATOD, BAEA and NDCS recommend stronger evidence, clearer definitions, robust due diligence on free tools and an explicit commitment that individual disability-related need remains the basis for DSA support.”
