Using AI to Create Image Descriptions in High Stakes Testing

Date & Time

Thursday, March 21, 2024 - 11:20 AM PDT  

Location

Elite 1-3  

Description

Assessment content containing images or graphics requires image descriptions for screen reader users. Construct measurement constraints around image descriptions poses a great challenge for designers of high stakes assessment whose aim is to provide an equitable and accessible test-taking experience to all its test takers; where there are elements of graphics that can be described and still preserve construct, users should have access to consistent descriptions. Without industry-wide standardized language for image descriptions, manual scripting for image descriptions is labor intensive, has an increased risk for error, and is not scalable. Working with Continual Engine and their Invicta™ product, ACT Inc. piloted a solution for automating image descriptions in its assessment content to generate consistent output with scalable results. Continual Engine designed and developed Invicta™, a generative artificial intelligence solution that uses sophisticated algorithms to accurately create detailed descriptions for graphics and images, which can be customized and edited to meet assessment requirements. In this session, ACT, Inc. and Continual Engine representatives will give an overview of the pilot, demonstrate the Invicta™ product used, review essential lessons learned throughout the pilot, and discuss possible future steps toward evolving an industry standard for image descriptions in high-stakes assessment.  

Audience

• Higher Education
• Information & Communications Technology
• Disability Specific
• K-12 Education
• Research & Development

Audience Level

Beginning  

Session Summary (Abstract)

Overview of ACT Inc. pilot using Continual Engine’s Invicta™, a proprietary generative AI solution, to generate image descriptions in high-stakes testing, discussing steps toward evolving an industry standard for image descriptions in assessment.  

Primary Topic

Artificial Intelligence (AI) & Machine Learning (ML)  

Secondary Topics

• Blind/Low Vision
• Digital Accessibility
• Education
• Product Demonstration

 

Session Type

General Track  

Presenters

• Allyson Vignola
  ACT, Inc.
• Vijayshree Vethantham
  Continual Engine US LLC

Add to Outlook/iCal (How it works)

Back to Session List