CAST 2026: "Accessibility, Automated Testing and AI - A Reality Check" with Chad Hester

This is a neat situation. We will have, effectively, back-to-back A11y talks. I'm giving a talk on Accessibility after lunch, mainly on how to focus on Accessibility with the entire software development life cycle. My goal is to give new information and, even more critically, NOT repeat too much of what the other talk is all about, so to that effect, I am here to both learn and filter my own presentation ;). 

Enough about me; the rest of this will be about Chad Hester, our presenter, and his talk about "Accessibility, Automated Testing and AI - a Reality Check".

There are many ways that we can say that we are "automating" accessibility testing. The nice thing about web accessibility especially is that there is a built-in vocabulary that expresses accessibility compliance. It's not always clear-cut or obvious, but some tags and attributes exist specifically in HTML to pinpoint and determine if accessibility features are being handled (or not). There's an entire framework for this (WCAG), and if you reference certain areas in the WCAG spec, you can determine if something is present. If it is, hey, compliant. If not, hey, we need to do something about that.

So to that effect, yes, Accessibility can be and is being automated. You can do it for free with the axe-core library (axe-core is created by Deque Systems). axe-core lets you call the library, read the web page elements, and based on what's there, it can tell you what is not compliant. It's a big hammer, but it can drive big nails. It can be finessed with extra code to get to individual items, too. I'm seeing something new this go around: BarrierBreak a11yInspect, so I want to see what that is. 

When “0 issues found” doesn’t mean accessible

I really appreciate Chad bringing this up early in the talk. Another thing to realize is that you can have a 100% compliant application, but it's miserable to use for everyone. Yes, everyone can use it, but how it's implemented may be truly sub-optimal. Automation will not tell you that. Humans need to make that judgement call. Automation can help find non-compliance, yes, but it can't tell you if your compliance is actually good or helpful.

WCAG is Ever Evolving

Often we look at WCAG as being definitive (think of it as the IEEE for accessibility features), but it's always evolving. To be most effective, we need to be aware of what version we want to be focusing our efforts on. WCAG 2.2 is the latest and greatest published version (as of 2023), but WCAG 3.0 has been under development and consideration for 10 years. Lots of cool things in it, but it's not authoritative or definitive, and things can and do change. It's a fun playground, and you can check out a variety of things, but if you want to be seen as doing the latest and greatest that's production valid, stick with WCAG 2.1.

Is AI an A11y?

Chad shared some examples of where AI was able to add value and what aspects they excel at. The WCAG rules, that's a no-brainer. Feed it the WCAG standard, and it can tell you if your code is compliant or not. That's a neat value add, don't get me wrong, but axe-core can do that for you too, with zero AI. My favorite statement, including when it comes to automation and AI, is "Tools Help, People Decide". AI in Accessibility will help with the mechanics, but it is no closer to making a true judgment call than automation by itself is.

Hello From Cocoa Beach: CAST 2026 is Underway: "How Avoiding Conflict Is Breaking Your Team"

 This is going to be an interesting and different CAST this go around. While I have live blogged conferences for years, I will not be doing much of that this go-around. That is because I am going to be doing live webcasting and video casting. We are launching a Webcast called "Responsive Ilities", we being me and Justin Bench. I will be posting the links to those as they go live. They will be fast clips, fun interactions, and we hope a good representation of the fun we are having here.

There will, of course, be some sessions we will not be vidcasting, such as our morning keynote right now. We are starting the conference with Dr. Jen Fry and her talk, "The Bugs Aren’t the Problem: How Avoiding Conflict Is Breaking Your Team"

The key message here is, "Software teams don’t struggle because they lack tools, frameworks, or technical skill. They struggle because the conversations that should happen… don’t."

I know this to be true, in some capacity, for every job I have ever had, both in my testing life and in the other things I do. I'm the enthusiastic, bubbly dude who likes doing, likes getting cool things done, and if there are conflicts or issues, we just let them roll off, and we figure we will just muscle through with our enthusiasm, and if we just kick enough butt, those issues will just disappear.

"Stop me, oh oh oh stop me.... stop me if you think that you've heard this one before!"

As you all know, I tend to riff on the ideas I'm hearing. I don't summarize the talk; I think through their points, and I take what I'm hearing and think through it with my own experiences.

I am intimately familiar with seeing something that I think might be wrong, but I'm unsure. Or I might want to make sure I can articulate everything before I vocalize my concerns.  Where does that come from? It comes from voicing concerns too early in the past, or thinking something is wrong but not having a full picture. What's the root problem here? I have an anxiety of looking like I don't know what I'm doing. Ever heard the phrase (paraphrased), "If you sit silently, you may be seen as a fool, but you can say what you are thinking and remove all doubt!" Yes, that's real. We all do that to a certain degree. We fear being seen as foolish or unwise, so the easiest tactic is to hold our thoughts until we are really sure. Maybe, possibly, we will gather what we need to make sure we have a thorough understanding and can present everything brilliantly. More often than not, though, life comes at you fast, and you are onto something else. Then when the problem rears its head, and others see it, you sit there thinking, "oooh, I knew that was an issue, but I didn't say anything because I didn't want to look stupid." Well, now the problem is out in the open, and I look more stupid because I held my peace until I was sure. 

Okay, so how can we counteract this very real issue people face? Part of it is practicing these words (again, this is how I handle this, or at least try to).

"I'm not fully sure I have a handle on this, but I'm noticing something that concerns me about [key area]."

"I may be missing something, but are you seeing an issue with [component]?"

"Are you familiar with this [area of misunderstanding] because I think I may have an issue... but again, I'm not sure."

Am I risking being seen as a fool right now? Absolutely. I'm sharing the fact that I'm "ignorant". Understand, that's a loaded word for many, and we hate hearing it about ourselves. Truth is, everyone is ignorant about a lot of things. Ignorance comes from Latin "ignorantia" and "ignorare", meaning "not to know" or "to be unacquainted with". I try to remind myself that "ignorance" is easily solved. I can learn, I can ask, and I can be taught. 

From a different place, I am often seen as the "psychic glue" in a group. That's a phrase a guitar player in one of my previous bands coined for me. He said that I had a tendency, due to my enthusiasm and general overall good nature (I think the phrase now would be termed 'Golden Retriever Energy' (GRE), a phrase that I actually love :) ), I make it possible for people who would struggle to interact a buffer, so that I cam the person in between who, because of my GRE, I can be that central glue point so we can all work together. Conflict doesn't stick to me or affect me when it's other people's conflict. If other people have issues with each other, I can be the go-between and easily smooth that challenge. The problem with GRE is that when the conflict comes to us, we can internalize it and think we are the problem. So we strive to make others happy and be cheerful, to hopefully cause the conflict to dissipate. I'm good at soothing issues. I'm not so good at actually resolving the conflict. The funny thing is that sometimes, my enthusiasm can have me jumping the gun at certain things, and then I get yelled at... but if people are yelling at me, they aren't yelling at each other. Since I have GRE, I can take having people be annoyed with me, and I can laugh or shrug it off easily. Net result: I glue everyone else together, and we function well,,, but you see the problem, right? Being psychic glue can be a good thing when it's applied correctly, but glue can also be messy if it's not applied correctly. Thus, I need to be willing to step back and speak more openly about the conflicts and be willing to not be seen as the happy-go-lucky golden retriever. I may not be able to be that glue at the moment, but I may be able to be the glue that is needed later on.

That's a great start. It's good to be back. Did you miss me :)?