Disabled citizens and the future of AI

Source: The post “Disabled citizens and the future of AI” has been created based on “Disabled citizens and the future of AI” published in “The Hindu” on 6th October 2026.

UPSC Syllabus: GS-2-Governance

Context: Artificial Intelligence (AI) is increasingly being used in government services, recruitment and healthcare. For persons with disabilities, it can provide greater independence, but its benefits will not automatically reach them unless accessibility and disability inclusion are made integral to AI development.

Opportunities of AI for Persons with Disabilities

1. Greater independence: AI can help persons with disabilities perform everyday tasks without depending entirely on others. For example, AI can help a visually impaired person read documents, understand photographs and access government forms.
2. Improved assistive technologies: AI can improve mobility by identifying accessible routes and can support communication through eye-tracking and voice-recognition technologies.
3. Better access to information and education: Speech-to-text software, automatically generated captions, image descriptions and digital assistants can make information more accessible.
4. Support in communication: AI can enable communication through technologies such as sign-language avatars and other assistive tools.
5. Support for independent living: AI-enabled systems can address individual needs and expand the possibilities for persons with disabilities to live independently.

Major concerns

1. Late inclusion: Persons with disabilities are being brought into the AI conversation only after AI systems are already being developed.
2. Poor accessibility: Many government websites and apps remain inaccessible to disabled citizens. Although the Rights of Persons with Disabilities Act, 2016 required accessible digital services from 2019, compliance remains an exception.
3. Weak implementation of accessibility rules: In the November 2024 Rajive Raturi judgment, the Supreme Court held that accessibility is a facet of the fundamental right to life and dignity. It also noted that accessibility rules had remained largely toothless.
4. Inherent AI bias: Research involving 21 language models showed that disability-related inputs made models more error-prone, negative in tone and more likely to stereotype.
5. Poor representation in datasets: The CLIP image model was 15 percentage points less accurate on photographs taken by blind and low-vision users. Objects important to disabled lives, such as white canes and Braille displays, appeared up to 17 times less often in training datasets.
6. Risk of discriminatory responses: AI systems may give inappropriate responses when users disclose their disability.
7. Lack of human fallback: A survey of 2,462 users of the NClude platform found that only 543 users completed previously closed tasks through AI alone, while others needed staff assistance.
8. Energy and infrastructure concerns: India’s growing data-centre capacity will increase electricity and grid requirements. Power cuts can seriously affect persons dependent on powered wheelchairs, oxygen machines and other essential equipment.

Way Forward

1. Every government AI system should be tested for disability bias.
2. Training datasets should contain genuine disability representation and should be gathered with consent.
3. AI systems should have human fallback mechanisms so that users are not left helpless when technology fails.
4. Government websites and apps must be made accessible instead of treating accessibility as an exception.
5. Data centres should adopt renewable energy and consider their impact on electricity grids and heat.
6. Persons with disabilities must have a voice in deciding the future of AI, including the data used to train models and the infrastructure supporting them.

Conclusion: AI should empower persons with disabilities rather than make them surrender existing tools in the name of technological progress. Their participation, genuine representation in datasets, accessible digital services and protection from disability bias are essential for an inclusive AI future.

Question: Artificial Intelligence has the potential to empower persons with disabilities, but its present development may also create new forms of exclusion. Discuss.

Source: The Hindu