The topic of artificial intelligence seems as ubiquitous as the weather in today’s ever-evolving world. The sentiment of these conversations tends to swing wildly between AI bringing about a long-awaited utopia, or the inevitable downfall of society as we know it.
Regardless of these polarizing discussions, the potential of AI should not be overlooked. After all, we are witnessing in real time how it is benefitting industries such as finance, healthcare, and even entertainment. Luckily for those of us in EdTech, AI is making waves (and automating tasks) on our doorsteps, too. An especially positive development is the impact on the notoriously complex task of making content accessible to as many users as possible. So, let’s consider five AI tools that you can use to make your courses not only engaging but also equitable.
ChatGPT exploded onto the digital landscape in November 2022 to much fanfare and has since been used to craft articles, finish homework, and even help you with your meme game. Â Fortunately for the ever-finicky process of accessibility, it is also a useful tool for keeping your alternative text (also known as alt text) as succinct as possible. It can identify redundancies and rein in the human tendency towards verbosity. Additionally, it can provide much-needed refuge from the tedious task of remediating auto-generated captions and creating transcripts by both formatting and including the necessary punctuation. Â However, even the mighty ChatGPT is not without its potential mistakes. It is vital that its responses are thoroughly checked to ensure that it has not deemed an important piece of information as irrelevant, or condensed your writing to nonsense. Always make sure your non-sighted user is getting all the relevant information they need from your content. ChatGPT offers both free and paid-for services.
SceneXplain could save significant time for learning designers, editors, and subject matter experts alike. What LD hasn’t spent hours googling the exact terminology to describe an image they’ve never seen before? This tool makes use of multimodal algorithms—algorithms that learn from varied data types, to provide complex responses. SceneXplain is, thus, capable of interpreting images and generating both captions and long visual descriptions. Even more impressively, it is also capable of interpreting videos and providing visual transcripts that would otherwise need to be manually created. Just think of the many emails and hours you’d save yourself and your SME! If you’re curious to see how SceneXplain outperforms its peers, you can take a look at this comparison image.
You will, however, need to proof these descriptions and transcripts to ensure that your learning objectives are being met. SceneXplain is a paid-for service only.
You’re three hours deep into making a PDF accessible. You finally put the last touches on your document, feeling confident that you’ve done everything in your power to make it accessible. Unfortunately, PDF Accessibility Checker vehemently disagrees. 🫨
CommonLook by Allyant is an AI-based tool that integrates with Adobe to automate the PDF remediation process. Importantly, CommonLook has been designed to remediate documents with both user experience and compliance in mind. This means that even users new to the process of PDF accessibility will be able to produce compliant documents without the technical knowledge PDF veterans have accumulated. Perhaps most impressively, Allyant has not lost sight of the need for the human touch in accessibility, noting that the tool itself is continuously evaluated by accessibility experts—with whom you can also speak.
In short, CommonLook can greatly expand your company’s capacity for PDF remediation, ensuring that human time is spent evaluating the user experience, rather than executing tedious tasks. CommonLook is a paid-for service only.
The world of online learning has become largely dominated by pre-recorded content such as videos—which are excellent tools for ensuring that the content is structured, well-edited, and accessible. However, synchronous learning still has an important place in education because it creates the space for direct contact with educators and the opportunity to learn collaboratively with other learners.
Ensuring that these synchronous sessions are fully accessible, however, is another obstacle to equitable learning. AssemblyAI could be the solution to this problem. It makes use of its own Conformer-2 model which has been trained with over a million hours of English audio. It can, thus, be used to convert live audio into accurate, fully punctuated text. This means that a lecture’s transcript will accurately reflect not only the content itself but also the educator’s interactions with the learners which is often where a lot of insight is created.
So far, we’ve mostly been exploring AI and accessibility from the designer’s perspective but what about tools that can make engaging with the content easier? Voiceitt is a speech-to-text tool that differentiates itself by focusing on users with speech disabilities, accents, or other characteristics that make being understood more difficult.
The tool is trained by each user to ensure personalized and accurate speech-to-text results. Much like AssemblyAI, Voiceitt is significant because it works in real-time, allowing users to actively participate in the moment. Voiceitt could be a significant tool for creating a truly equitable educational experience, ensuring that all questions get asked and all learners can participate fully.
Returning to the designer’s perspective, a tool like Voiceitt could expand your company’s talent pool by providing the framework to work with people whose speech patterns may have been a barrier to entry previously.
AI promises to revolutionize many industries, including Ed Tech, however, it is important to note that accessibility is inherently a human experience which means that accessibility will always require the human touch.Â
As you have seen with most of the tools we have explored, AI can provide automation and frameworks, but the efficacy of these tools has to be assessed by a human to ensure that they are functioning as intended. This means that those of us in EdTech need to stay abreast of the tools that are evolving and leverage what is available whilst maintaining our knowledge of accessibility and its associated processes.
Because when all is said and done (or prompted), we are responsible for capturing and conveying the messages of our content to everyone.
This blog post is made available by the author for educational purposes only and to provide general information. All views expressed are the author’s own and do not represent the opinions of any entity whatsoever, to which they have been, are now, or will be affiliated. If you have a specific problem related to this topic and need advice, contact Construct Education directly.
© 2024 All rights reserved
Construct Education is the trading name of Proversity.Org Ltd (UK entity), Proversity Inc (USA entity) and Proversity Org Incorporated in UK (South African entity)
