One of the better known epilepsy accessibility frameworks today is the warning displayed before media content that features heavy flashing lights. In Improving Online Accessibility for Individuals with Photosensitive Epilepsy (2022), South asks: what do we do about the prevalence of similar effects in the new media era? Writing ahead of a conference, South structured a paper by describing her pre-existing work on accessibility and then put forth her future proposals.
PhotosensitivityPal: Consumer-Driven Design Justice
When South conducted interviews to identify gaps in online accessibility, those with epilepsy named a variety of actions they use to respond to triggers. They ask friends to check for dangerous content, rely on creators themselves, or even throw their phone away from their person (pp. 2-3). In response, South developed a browser extension named PhotosensitivityPal that can scan for seizure-inducing content in GIFs. This is, as she put it, a consumer-driven system: the end user has control over what content is blocked, rather than relying on platforms or content creators (p. 2). This practice is reminiscent of what Sasha Costanza-Chock writes about in Design Justice: Community-Led Practices to Build the Worlds We Need. In this book, Costanza-Chock points to how specific communities challenge digital structural inequality (“specifically, those who are intersectionally disadvantaged or multiply burdened under the matrix of domination (white supremacist heteropatriarchy, ableism, capitalism, and settler colonialism)”). As the abstract notes, Costanza-Chock’s book is meant to address intersectional perspectives, so reducing it to accessibility does not do the writing justice. However, a key takeaway is that there are particular values embedded in the systems we use. This occurs by way of platform affordances (the design-enabled tasks you can perform on any given piece of technology or platform), the people that build them, what we privilege in design, and the ways that we teach design (pp. 23-24). It becomes the responsibility of disadvantaged communities to problem-solve and create digital technologies or environments that address their needs (pp. 169-170, 195). While it is noted that the Americans with Disabilities Act extends to web design, compliance is viewed as a “burden” by designers (p. 223). It should go without saying that this perspective is a perpetuation of prejudice and stigma.
In Chapter 4, “Design Sites: Hackerspaces, hackathons, fablabs, and DiscoTechs”, Costanza-Chock discusses design loci themselves (p. 135). These are all communities where coding and design have been made more accessible – not exclusively in the functional sense, but always about representation. Here, communities step in to create environments that reflect their respective demographic, ideally not reproducing the matrix of domination (many aspire to be “inclusive” but fail to dissect that principle – pp. 155-156). Though Costanza-Chock flags these sites as failing to challenge the matrix of domination, she points to a communal view of autonomy, empowerment, and a hopeful future where “shared spaces are no longer needed” (p. 158). She explains that “democratizing access to design tools and skills is truly important”, even though a critical analysis might expose flaws (such as when the spaces return to being dominated by white, middle-class, cisgender, abled men – pp. 159, 161-162). For our context, we should read this work to understand why these cultures emerge (p. 165). South developed her own tool, stepping in when tech giants failed.
Though South was able to develop a tool that mediated some of the challenges experienced by those with epilepsy, she recognized that there are still flaws in the online ecosystem. These concerns take shape in a few different ways. First, advances in technology pose continual challenges – South uses the example of virtual reality (VR), where the medium encompasses the entirety of one’s field of vision (p. 3). Likewise, the Epilepsy Foundation offers tips on managing triggers that stem from auto-playing videos. Social media companies want to keep you on their applications, and as they place more stock in short-form, attention-grabbing videos, we can see where this might lead to accessibility issues (n.d.). Secondly, malicious intent has always been a problem with the use of GIFs in targeted attacks dating back to 2008 (pp. 1-2). This was how South narrowed the scope of her work, but it remained a concern. Lastly, South was interested in developing PhotosensitivityPal to exceed boundaries set by users. (p. 3). Her logic was that doing so would provide an extra layer of protection against these accidental or malicious exposures.
Reflecting on the Experiences of Other Communities: Captioning
The question remains: Are creators always conscious of the need for these warnings (or will they ignore them out of malice – p. 3)? Can systems be widely implemented that flag videos for potential triggers? As demonstrated by South’s work, innovative technology such as PhotosensitivityPal is important, but there should be a desire to make it widely available and effective. We can use captioning technology in today’s digital age as a comparison to underscore the importance of quality technology that is efficient and accessible for everyone.
Closed captioning has been implemented across many digital platforms with video. This is impactful for those that are hard of hearing. Unfortunately, flaws in the technology can limit comprehension (Smith et al., 2017). In Smith et al.’s Reading Between the Lines: Accessing Information via YouTube’s Automatic Captioning study (2017), they explain that automatic captioning uses speech-recognition technology that relies on probability and statistics (p. 115). As such, we have to understand that at the end of the day, a model is still a model. It can struggle with accents, context, and other variables that make these assumptions inapplicable (p. 116). To see how well YouTube’s auto-captioning system functioned, 75 participants were asked to watch middle-school science videos that had been categorized as having “few” errors or “high” errors, then answer comprehension questions. Their participants were separated into groups, allowing Smith et al. to examine the effects of each variable: 1.) sound without captioning, 2.) sound with automatic captioning, or 3.), automatic captioning without sound (p. 119). What Smith et al. discovered was that when auto-captions were accurate, the video’s message was relayed effectively – sound or no sound. Inaccuracies led to misunderstandings, even with “hearing, college-educated adult readers”. They could not infer the message from videos that had high errors and no sound. As Smith et al. write, “expecting school-age students who are d[/D]eaf/hard of hearing with varied reading abilities to perform any better would be inappropriate […]” (p. 124). Essentially, the accessibility tool exists, but it is ineffective if it is not moderated or used to begin with.
In Perspectives of Deaf and Hard of Hearing Viewers of Captions, Janine Butler put together a focus group of 20 college students. The study covered a range of topics related to captions (p. 536). When asked about digital captioning, participants offered both criticism and praise, with some giving credit towards improvement in recent years (p. 544). Participants called for further technological advancement, and urged creators themselves to be responsible for captions (p. 545). In concluding this portion of the paper, Butler writes:
“To nurture the continual growth of captioning in live and social media contexts, yet more awareness is needed so that video creators include captions in videos they put online.” (p.545).
This emphasis on advocating for creators to provide accessible content mirrors that of navigating safe environments for epilepsy. Across both scenarios, design and education have been recognized as important to creating equitable online experiences.
A Brief Conclusion
As digital accessibility takes shape for epilepsy, we can reflect on what questions need to be asked of creators, platform developers, and our community itself. As Laura South discovered, photosensitive individuals were taking drastic measures to protect themselves online. Her work to remedy this was ongoing and she even tried to anticipate further issues. When writers talk about design justice, they recognize that diverse coding communities emerge when their needs have not been addressed by major developers. In separate accessibility case studies, we learn that progress can be made, but there is generally an ongoing battle between who is responsible and where technology continues to fall short. Trying to untangle all of these concepts is not easy, but it is important. We need a world where people do not have to be cautious about how they surf the Web.
Butler, J. (2019). Perspectives of Deaf and Hard of Hearing Viewers of Captions. American
Annals of the Deaf (Washington, D.C. 1886), 163(5), pp. 536, 544-545. https://doi.org/10.1353/aad.2019.0002
Costanza-Chock, S. (2020). Design Justice: Community-led practices to build the worlds we
need. MIT Press. https://doi.org/10.7551/mitpress/12255.001.0001
Epilepsy Foundation. (n.d.). Photosensitivity and seizures.
[Reviewed by Elaine Wirrell MD.] https://www.epilepsy.com/what-is-epilepsy/seizure-triggers/photosensitivity
Smith, C., Allman, T., & Crocker, S. (2017). Reading Between the Lines: Accessing Information
via YouTube’s Automatic Captioning. Online Learning (Newburyport, Mass.), 21(1), 115-124. https://doi.org/10.24059/olj.v21i1.823
South, L. (2022). Improving Online Accessibility for Individuals with Photosensitive Epilepsy.
CHI Conference on Human Factors in Computing Systems Extended Abstracts, 1–4. https://doi.org/10.1145/3491101.3503805
