DCMP Includes AI Scene Description Tool for Students with Visual Disabilities

The federally funded Described and Captioned Media Program (DCMP) has developed an "AI Scene Description Tool" add-on to its video player to increase accessibility to video content for blind students and those with a low vision disability. The tool is live and is currently in beta testing.

The tool uses artificial intelligence (AI) to "describe the visual elements of any frame within a paused video," the organization said.

This tool, a supplement to the program rather than a replacement ("DCMP will continue to produce audio description with talented writers and voicers," the organization said) creates a secondary audio track that describes visual information in an educational video. The student presses the "AI Scene Description" button to pause a video and ask detailed questions of the AI. Any paused scene is described via text-to-speech and transcribed.

This is especially helpful when there is "'quiet time' in a video, where the additional narration is inserted," DCMP added.

AI-generated scene descriptions rely "strongly on human-created metadata for each video," DCMP said, and "guardrails" are in place to keep AI responses "contextual to the video" using grade-appropriate vocabulary, educational standards, captions, and audio description. Students cannot have a conversation with the AI. User data is protected and never sent to the AI model, the organization said.

The DCMP program is available with free registration to any teacher or family with at least one early learner through Grade 12 disabled student. Each video has a "Provide Beta Feedback" button to share insights on how it's working for learners.

Teachers or students log in at dcmp.org and select the "beta AI Scene Description button under any video. Teachers give access to students by going to "Account > Students," and selecting "Allow students to use AI Scene Description."

DCMP is funded by the U.S. Department of Education under its Office of Special Education Programs (OSEP), and administered by the nonprofit National Association of the Deaf (NAD).

Visit the DCMP registration page to apply for a free membership and get access to its many educational resources for students who are deaf, blind, hard of hearing, visually impaired, or deaf-blind, and professional development resources for teachers, families, and other caregivers.

About the Author

Kate Lucariello is a former newspaper editor, EAST Lab high school teacher and college English teacher.

Featured

  • A young person sitting on a couch in a bright living room during the daytime, holding a tablet, with sunlight streaming through large windows and pastel-toned furniture.

    Balancing Screen Time and Student Wellness

    Student mental health is in crisis, and excessive screen time is a significant factor. Here are four ways to help students find and maintain a healthy balance with technology.

  • abstract circuit patterns, featuring small icons of coins and dollar symbols in blue, green, and gold tones

    Report: 90% of Schools Depend on E-rate Funding Every Year

    A new report from Funds for Learning highlights the critical role of the Federal Communications Commission's E-rate program in funding essential technologies and services for K-12 schools.

  • interconnected geometric human figures forming a network

    CoSN: School Staffing Is the Top Hurdle to K-12 Innovation

    Hiring and keeping educators and IT staff remains the top challenge for K-12 education in 2025, according to the latest Driving K-12 Innovation Report from the Consortium for School Networking (CoSN).

  • computer with a red warning icon on its screen, surrounded by digital grids, glowing neural network patterns, and a holographic brain

    Report Highlights Security Concerns of Open Source AI

    In these days of rampant ransomware and other cybersecurity exploits, security is paramount to both proprietary and open source AI approaches — and here the open source movement might be susceptible to some inherent drawbacks, such as use of possibly insecure code from unknown sources.