AI in the Workplace: Assisting Blind and Low Vision Professionals
This content is also available in:
Español (Spanish)
For many people, artificial intelligence (AI) only entered the spotlight with the public release of ChatGPT in late 2022. But for blind and low vision professionals, AI has been quietly transforming accessibility and independence for years through tools like optical character recognition apps, screen-reading assistants, and automated description services.
As someone who has been totally blind since birth and works as a rehabilitation counselor, I’ve long viewed AI as a key to my independence. It started with apps that finally allowed me to read my mail without assistance. I’ve always been fascinated by how these technologies could enhance my independence, empowering me to live on my own terms.
Today, with AI rapidly becoming part of everyday workplaces, the key question is: how can we ensure this technology is used to create more accessible, inclusive, and equitable opportunities at work?
The Rise of AI and What It Means for Accessibility
AI has unequivocally stepped into our everyday workflows. Before ChatGPT, the idea of AI assisting with daily tasks felt futuristic; now, it’s becoming a practical reality for millions. For people who are blind or low vision this shift is particularly exciting. We’ve long sought practical solutions to bridge accessibility gaps, and the advancements in AI, especially in areas like natural language processing and advanced OCR, represent what many are calling the “new frontier” for digital inclusion.
With AI technology now no longer in the shadows, these tools can be leveraged in ways that will enhance the productivity of all employees, including those of us who are blind or low vision. AI capabilities go way beyond mere text analysis and generation. It has since developed capabilities including image and video creation, text-to-speech with expressive voices, coding, and performing research on the Web. The promise is clear: AI could simplify how we interact with information, navigate environments, and perform tasks, ultimately creating truly equitable workplaces.
AI in the Workplace: Benefits
The benefits of AI in professional settings are substantial. Enhanced productivity can level the playing field in ways we’ve never experienced before.
The following are examples of AI boosting accessibility and efficiency on the job:
- Reading and summarizing documents – Quickly process lengthy PDFs, reports, or meeting notes that may not be fully accessible.
- Formatting reports and case notes – Fit raw notes into templates or produce structured summaries that meet workplace standards.
- Scheduling assistance – Use AI to propose meeting times, manage calendars, or create reminders.
- Preparing presentations – Create outlines, talking points, or alt text for images in slides.
- Navigating inaccessible software – Leverage AI screen-reading or image-recognition tools to access buttons, forms, or menus that don’t work with traditional screen readers.
- Interpreting visuals – Get AI-generated descriptions of charts, graphs, or images used in reports, trainings, or presentations.
- Proofreading and clarity checks – Run drafts through AI for grammar, readability, and formatting suggestions.
- Data entry support – Use AI to clean up or structure information for or from spreadsheets or databases.
- Meeting support – Generate real-time transcripts, summaries, or action item lists from audio recordings
AI in the Workplace: Pitfalls and Risk
However, employers are understandably skeptical about allowing their employees to use AI tools in the workplace for several reasons. First and foremost, many employers handle confidential information, and they have legitimate concerns about third parties accessing this information. AI inputs are often stored, reviewed, and used for training models, which poses a serious security risk for companies that must safeguard data.
Employers are also concerned with hallucinations in which an AI will make up information rather than plainly state that the information being sought is unavailable. There have been several instances when lawyers have been reprimanded because they used fictitious legal citations fabricated by ChatGPT. In some cases, such as when making a medical diagnosis, writing reports based on test score outcomes, or recommendations based on a consumer’s needs, hallucinations could prove to be drastically wrong.
For blind and low vision individuals who may be using AI for reading materials that are not accessible-stand-in for inaccessible information, this risk can be particularly problematic.
It is important for anyone with a real desire to use AI in their workplace to address these concerns and outline how they plan to deal with them. For instance, using a completely offline AI model or opting out of model training to ensure information stays private, and having all AI content reviewed and edited before using it. The key is implementing a human-in-the-loop verification system to catch potential errors before they become problems.
Requesting AI as a Workplace Accommodation
When I initially requested to use AI on the job, it was to address several issues I was having. As a rehabilitation counselor, I found writing intake summaries and case notes challenging because they are formulaic and require a template. I often find the task to be incredibly difficult due to the amount of focus and attention to detail required. I found that AI could help me get organized, fit my raw notes into a template, and enable me to complete the case work required for my position.
When I initially asked for accommodations, there was no formal system, and I was verbally advised that I could use AI if I was able to access it on the network. I ended up using Olympia, a company that provides virtual employees. Because it was an informal arrangement, I had to pay for the program myself, and I still felt wary about using the tool, mainly because it could get blocked without warning, and I might be liable if any data was leaked.
The key to successfully requesting AI as a workplace accommodation is framing it as assistive technology under the ADA to quell any fears that it may be mistaken for a luxury. Here are essential talking points when making your request:
- Frame it as assistive technology: Position AI tools as the digital equivalent of screen readers or other established accommodations.
- Address security concerns proactively: Propose using offline models or enterprise versions with data protection.
- Provide specific use cases: Explain exactly how AI will help you perform essential job functions.
- Offer verification processes: Suggest human review systems to address accuracy concerns, either you or a clerical staff member.
- Present cost-benefit analysis: Show how AI accommodation could increase your productivity and reduce other support needs.
Thankfully, my organization has become more progressive about AI’s practical applications at work, and I hear they plan to install Microsoft’s Copilot onto our systems with an upgrade later this year, enabling everyone to benefit from the technology.
Aside from ChatGPT, numerous AI tools are helpful in the workplace, with some being more beneficial in specific industries. The primary tools I use can be grouped into two broad categories: information access and content creation.
Information Access Tools
BeMyAI (Free): Part of BeMyEyes, this service allows blind and visually impaired individuals to connect with sighted volunteers for help using the phone’s camera. BeMyAI is the automated arm of this service, which describes images on the phone, can explain what’s on the computer screen, and allows a person to use the camera to describe their physical environment. For BVI professionals, this is invaluable for understanding visual content in presentations, documents, or workplace environments.
EnVision AI and Seeing AI (Free): These tools enable users to receive feedback from the camera, featuring modes such as currency reader, color and light detection, and facial recognition. For quick tasks such as reading a business card or finding out what my oven temperature is, the quick reading mode is faster than BeMyAI, as there seems to be no online processing of the data for results to show up. In workplace settings, these are perfect for reading printed materials, labels, or signage.
ViewPoint (Free): A recent discovery made by Nibble Nerds leverages Google’s Gemini AI to read the computer screen on a Windows PC, acting as an augmentation to any screenreader. It uses optical character and image recognition to allow a person to click on elements that aren’t typically accessible, and can describe pictures and read PDFs that might not be easy to read under normal circumstances. This tool is particularly valuable for navigating inaccessible workplace software or websites.
Content Creation Tools
ChatGPT, Claude, and Gemini (Free tiers available): These general-purpose AI assistants excel at workplace writing, summarizing lengthy documents, composing professional emails, and formatting reports. They’re particularly useful for BVI professionals who need to process and respond to large volumes of text-based information quickly.
ElevenReader, Speechify, NaturalReader, Peach AI, and SpeechCentral (Paid with the exception of SpeechCentral, which is free if you use VoiceOver): Provides high-quality text-to-speech for documents and webpages, making it easier to consume written content efficiently. The natural-sounding voices can help with fatigue that comes from extended screen reader use.
Looking Ahead: How AI Might Reshape Accessibility
The future of AI in accessibility looks incredibly promising. We’re moving toward AI agents that can act as navigators, automatically fixing or translating inaccessible web content—imagine AI that can instantly add alt text to unlabeled images or describe poorly formatted documents in real-time.
Image generation tools are beginning to incorporate accessibility from the ground up, creating visuals with described outputs that make visual content inherently more accessible. For multimedia, we’re seeing rapid advancement in transcription services like Otter.ai and Whisper for meetings, audio summarizers that can distill hours of content into key points, and video description generators like PiccyBot that can provide real-time narration of visual content.
The key to realizing this potential is ensuring that accessible-first design principles are built into these tools from the beginning, rather than retroactively as an afterthought. As AI becomes more sophisticated, the opportunity to create truly inclusive digital experiences grows exponentially.
Advocacy, Awareness, and Adoption
The integration of AI into workplace accessibility represents more than just technological advancement—it’s a fundamental shift toward digital inclusion. For BVI professionals, these tools aren’t just nice-to-haves; they’re career-changing technologies that can unlock new opportunities and level professional playing fields.
The importance of including BVI users in AI tool design cannot be overstated. Our unique perspectives and needs drive innovation that benefits everyone. I encourage all professionals to advocate for access to AI at work, share your experiences with these tools, and actively request support from your organizations.
The future of work is increasingly AI-augmented, and we must ensure this future includes everyone. By exploring these tools, requesting accommodations, and sharing our successes and challenges, we can help shape an AI-powered workplace that truly works for all.
Resources