Contributor(s)

David Fazio, Founder/CEO, Helix Opportunity LLC

Description

Helix Opportunity has developed a browser extension, Ottitor, that is available on all chromium-based browsers and utilizes Large Language Models, Natural Language Processing, Optical Character Recognition, and other advanced technologies, to analyze web content to identify more complex accessibility issues that typically require manual intervention.

It has been designed with a streamlined, point of use user interface that reduces distractions and fatigue, enhances focus, and pesents easy use.

Many Web Content Accessibility Guideline (WCAG) rules necessitate a semantic understanding of content, which traditionally could only be verified through human intervention. This process is often time-consuming and prone to inconsistencies. In developing Ottitor, we explored leveraging the capabilities of artificial intelligence to automate these checks, making accessibility audits more efficient and reliable. This digital accessibility auditing tool has been successfully implemented as a browser extension, allowing for seamless integration into the web browsing experience. It effectively addresses various WCAG rules that require nuanced comprehension, ensuring that digital content is accessible to all users without the extensive manual effort typically required.

One of the significant features of Ottitor is its ability to address the challenges posed by abbreviations, idioms, and jargon. AI capabilities make it possible for Ottitor to analyse links for coherence, as well . Our AI implementation expands abbreviations for better comprehension but also provides interpretations in multiple languages, making information accessible to a broader audience. Idioms and jargon can be tricky for users, especially those unfamiliar with a language. Ottitor’s AI implementation also provides the meaning of terminologies and explanations for idioms and phrases. It supports multiple languages to ensure inclusivity.

Ottitor had to overcome several limitations with existing AI models. For example, very complex or heavy web pages had to be procesed through multiple scans due to token limitations or exceeding the number of allowable API requests. These constraints can impact the tool’s ability to analyze and provide feedback on large-scale content effectively. Despite these challenges, we remain committed to refining our technology and expanding its capabilities based on user feedback and continuous development.

We are currently exploring advanced features like screen recording analysis, video and audio file analysis, image analysis, and dynamic content analysis. By recognizing patterns and relationships in data, we aim to enhance the auditing process even further, ensuring comprehensive accessibility reviews.

Presenter(s)

David Fazio