WEBVTT 00:00.000 --> 00:09.400 I love a warm welcome in this excellent. 00:09.400 --> 00:12.480 And also, I have stickers at the end, especially for people who 00:12.480 --> 00:15.680 clap at, if I make jokes or anything about our stickers, yes, 00:15.680 --> 00:18.640 to get ready to read. 00:18.640 --> 00:22.840 So my target, sorry, my name is Mike Gifford. 00:22.840 --> 00:26.040 I'm an open standards and practices lead at civic actions 00:26.040 --> 00:29.960 where a US-based company, I'm a Canadian 00:29.960 --> 00:32.200 live in Ottawa, Canada. 00:32.200 --> 00:33.720 We do a lot of work with the government. 00:33.720 --> 00:35.720 I'm also a WC3 invited expert. 00:35.720 --> 00:39.360 And I work with folks like HITTY on all kinds of stuff 00:39.360 --> 00:42.040 at the WC3 and really have an opportunity 00:42.040 --> 00:44.640 to learn from experts. 00:44.640 --> 00:47.360 And I came to this to accessibility 00:47.360 --> 00:49.080 from the Drupal community. 00:49.080 --> 00:52.440 And Drupal is a large content management system. 00:52.440 --> 00:54.840 And I've been spearheading the accessibility 00:54.840 --> 00:58.880 initiatives in Drupal for, I don't know, since 2009. 00:58.880 --> 01:01.360 So I don't want to do the math. 01:01.360 --> 01:03.840 But I've learned a lot through the process. 01:03.840 --> 01:08.720 And we were also for ATHG, the offering to accessibility 01:08.720 --> 01:09.280 guidelines. 01:09.280 --> 01:12.520 We were one of the two test cases that were acquired 01:12.520 --> 01:15.440 in order to get that standard into place. 01:15.440 --> 01:17.040 It's been a really interesting opportunity 01:17.040 --> 01:21.600 to learn about how to go off into incorporate 01:21.600 --> 01:24.200 accessibility into offering tools. 01:24.200 --> 01:26.440 And so I wanted to walk through some of those ideas 01:26.440 --> 01:28.600 and also look ahead at where we're going. 01:29.160 --> 01:30.960 There's a lot of changes right now for authors 01:30.960 --> 01:33.640 as for our VWNLs. 01:33.640 --> 01:38.440 So right now, we're doing, we're doing some 01:38.440 --> 01:39.640 as a little bit insane. 01:39.640 --> 01:42.440 We're not really expecting for some reason 01:42.440 --> 01:46.040 authors to become accessibility experts. 01:46.040 --> 01:47.360 And so there's a cycle of failure. 01:47.360 --> 01:50.560 You'll look at all of the, I am also authorizing 01:50.560 --> 01:53.640 that one of the authors of the Web Almanac this year, 01:53.640 --> 01:55.720 the accessibility chapter. 01:55.720 --> 01:57.760 And as part of the Web Almanac, we're 01:57.760 --> 01:59.800 trying to evaluate how much of the Web 01:59.800 --> 02:02.040 have we being, how does it look, how accessible 02:02.040 --> 02:03.960 has it compared to last year? 02:03.960 --> 02:06.840 And despite things like the European Accessibility Act 02:06.840 --> 02:09.600 coming into effect this past year, 02:09.600 --> 02:11.160 we haven't seen a great improvement 02:11.160 --> 02:16.120 over the accessibility in Europe or many other places. 02:16.120 --> 02:17.520 Strangely the one place where we have 02:17.520 --> 02:21.120 seen improvements in accessibility, the dot AI domain 02:21.120 --> 02:23.360 is now the third most accessible top level 02:23.360 --> 02:26.280 domain in the world in the surveys that we've done. 02:26.280 --> 02:29.040 So there might be something there. 02:29.040 --> 02:33.440 And we need to train authors because they change over time 02:33.440 --> 02:36.520 or they have other subject matter experts in other areas. 02:36.520 --> 02:38.440 They're not trained or have the time 02:38.440 --> 02:40.000 to be accessibility experts. 02:40.000 --> 02:42.480 And we should expect them to. 02:42.480 --> 02:46.280 We need to be starting to use these tools as filters 02:46.280 --> 02:48.960 to help authors create better content. 02:48.960 --> 02:51.600 We need to be more strategic about how we do that 02:51.600 --> 02:56.080 so that we're, and shift the focus from authors 02:56.080 --> 03:02.480 to authors, because right now we're all of the energy, 03:02.480 --> 03:05.120 all of the focus, all of the money in the accessibility 03:05.120 --> 03:05.640 industry. 03:05.640 --> 03:07.360 And there is a lot of money that isn't 03:07.360 --> 03:09.560 is going into the accessibility industry. 03:09.560 --> 03:11.160 It goes into auditing. 03:11.160 --> 03:13.680 And that's not a good space to fix the problems. 03:13.680 --> 03:14.880 If you wanted to fix the problems, 03:14.880 --> 03:17.520 we need to be looking upstream, looking at fixes 03:17.520 --> 03:21.040 to open source tools, like Drupal and others, 03:21.040 --> 03:25.040 and not focusing on auditing. 03:25.040 --> 03:28.960 So it's not really a missing standard ATG is out there. 03:28.960 --> 03:30.640 It's just not applied. 03:30.640 --> 03:33.640 So many people who have heard about the web content 03:33.640 --> 03:36.160 accessibility guidelines, they may have heard about 03:36.160 --> 03:38.640 the evolution of those guidelines 03:38.640 --> 03:44.040 and how we're now moving towards the WKG 3.0, which is great. 03:44.040 --> 03:46.520 And there's lots of interesting stuff that's happening there. 03:46.520 --> 03:50.800 There might be a WKG 2.4, who knows. 03:50.800 --> 03:52.440 Certainly we have a, or sorry, 2.3. 03:52.440 --> 03:54.440 There's certainly a 2.2 right now that will probably 03:54.440 --> 03:58.680 be coming into law in the, in the Ian 3154 03:58.680 --> 04:02.200 and our 9th standard in the next year, I hope. 04:02.200 --> 04:06.080 But so often organizations overlook authors 04:06.080 --> 04:09.320 and look over the people who use their interface. 04:09.320 --> 04:12.200 And in Drupal, we've been incorporating 04:12.200 --> 04:14.920 and thinking about authors since Drupal 7, 04:14.920 --> 04:19.240 which is a little while ago, because it's reached the end of life. 04:19.240 --> 04:22.080 But most people don't think about the 25% of the population 04:22.080 --> 04:23.840 that has a disability. 04:23.840 --> 04:26.600 And we're not, you know, that includes authors. 04:26.600 --> 04:28.560 It's not just the general public. 04:28.560 --> 04:31.320 If there's government employees or university employees 04:31.320 --> 04:33.800 or using these interfaces to provide content, 04:33.800 --> 04:37.960 that includes that 25% of the population. 04:40.600 --> 04:43.640 Part B of A tag is often completely forgotten. 04:43.640 --> 04:45.200 Like sometimes people will think about part A 04:45.200 --> 04:47.200 trying to make the back end accessible. 04:47.200 --> 04:49.760 But they won't often think about part B, which is, 04:49.760 --> 04:52.000 how do we try and support those authors? 04:52.000 --> 04:53.800 How do we build in systems? 04:53.800 --> 04:57.720 So this easier for authors to do the right thing. 04:57.720 --> 04:59.480 And not to suddenly assume that those authors 04:59.480 --> 05:02.320 are going to be accessibility experts. 05:02.320 --> 05:03.480 And so there's things that we've done, 05:03.480 --> 05:05.280 like in Drupal 7 or sorry, Drupal 8, 05:05.280 --> 05:08.400 we work to, when we tried to address A tag, 05:08.400 --> 05:12.600 we did things to make sure that all text was required 05:12.600 --> 05:13.840 on images. 05:13.840 --> 05:17.800 And we were going to, one of the first CMSs to do that. 05:17.800 --> 05:20.240 And it was a really excellent opportunity 05:20.240 --> 05:25.840 to try and take this material and try and say, 05:25.840 --> 05:30.240 how do we apply it in the context of our content management 05:30.240 --> 05:30.640 system? 05:34.840 --> 05:40.080 Now, A tag 2 is an decade old. 05:40.080 --> 05:41.760 It hasn't been applied all that often 05:41.760 --> 05:44.360 to a lot of content management systems 05:44.360 --> 05:46.360 because it hasn't been required. 05:46.360 --> 05:49.120 It's not part of contracting in general. 05:49.120 --> 05:52.600 It's not part of legislation in general. 05:52.600 --> 05:59.600 And now we have a rush towards adoptions of LLMs 05:59.600 --> 06:01.560 and that's adding the additional pressure 06:01.560 --> 06:04.880 to making things more accessible. 06:04.880 --> 06:08.320 So how do we try and end this process? 06:08.320 --> 06:11.680 Get move from from having guardrails, 06:11.680 --> 06:14.240 like error messages that pop up that say, 06:14.240 --> 06:17.240 oh, this must be like, you can add error messages 06:17.240 --> 06:19.360 and validation errors to sort of force authors 06:19.360 --> 06:21.320 to do things, but it would be a lot better 06:21.320 --> 06:24.800 if we could provide better guidance, provide better tools 06:24.800 --> 06:27.400 in line so that we're thinking about accessibility 06:27.400 --> 06:30.040 like a spell check. 06:30.040 --> 06:34.760 We would all, if we took away spell check right now, 06:34.760 --> 06:36.480 all of us would be worse writers 06:36.480 --> 06:38.720 because we expect to have the spell check there 06:38.720 --> 06:39.920 and grammar check. 06:39.920 --> 06:41.480 Like these are things that are built in 06:41.480 --> 06:43.680 that support our communication. 06:43.680 --> 06:47.200 We don't have that yet for accessibility. 06:47.200 --> 06:49.480 Even though there's tools built in 06:49.480 --> 06:51.600 to do accessibility in a lot of places, 06:51.600 --> 06:54.000 nobody includes them by defaults. 06:54.000 --> 06:56.160 And we need to start changing that. 06:56.160 --> 06:58.440 Julia actually is one of the few CMSs 06:58.440 --> 07:02.120 that includes Sally by default. 07:02.120 --> 07:04.600 Sally will come up later. 07:04.600 --> 07:09.640 But there's a, it's trying to, 07:09.640 --> 07:11.080 to think about how do we support 07:11.080 --> 07:12.640 the authors and the creation of content 07:12.640 --> 07:13.960 is a real challenge. 07:14.920 --> 07:17.080 But AI also gives us the ability 07:17.080 --> 07:20.160 to provide more customized information. 07:20.160 --> 07:23.880 So, do we want to provide information 07:23.880 --> 07:26.040 that's for different audiences? 07:26.040 --> 07:27.720 Do we want to provide a plain text version 07:27.720 --> 07:29.280 and easy language version? 07:29.280 --> 07:30.880 Do you want to provide it in different languages 07:30.880 --> 07:31.720 for different roles? 07:31.720 --> 07:34.040 Like we can do that now. 07:34.040 --> 07:35.600 And a lot of that is actually getting built 07:35.600 --> 07:37.280 into the browser. 07:37.280 --> 07:38.880 There's AI in the browser, 07:38.880 --> 07:42.680 which will provide some interesting opportunities 07:42.680 --> 07:46.160 to go off into to customize and personalize information. 07:46.160 --> 07:48.200 It will also introduce a great many problems. 07:48.200 --> 07:51.440 So, by no means as AI is a panacea 07:51.440 --> 07:52.720 that will solve all of our problems 07:52.720 --> 07:54.360 that is both going to give us more problems 07:54.360 --> 07:56.440 and more solutions at a faster rate 07:56.440 --> 07:57.720 than we've ever seen before. 08:00.520 --> 08:02.200 So, about four years ago, 08:02.200 --> 08:04.440 the Funka and Sweden went off 08:04.440 --> 08:06.640 and let an international study in preparation 08:06.640 --> 08:11.240 for the European accessibility act 08:11.240 --> 08:13.840 to try and get the content management systems 08:13.840 --> 08:15.440 that were popular in Europe 08:15.440 --> 08:20.440 to start thinking about improving their offering and interface. 08:20.440 --> 08:21.760 And there's a bunch of organizations 08:21.760 --> 08:23.520 that came and were a part of that. 08:23.520 --> 08:25.640 And he didn't want to take the A-take process 08:25.640 --> 08:26.960 and just reuse it. 08:26.960 --> 08:30.880 They wanted to be able to look at finding out 08:30.880 --> 08:32.200 what are the practical tools, 08:32.200 --> 08:34.760 four CM authors that we beneficial for them. 08:34.760 --> 08:37.320 And there's a bunch of stuff that came out of that. 08:38.200 --> 08:40.000 And there were several CMS authors 08:40.000 --> 08:41.880 and editor editing projects were involved. 08:41.880 --> 08:48.240 Like, we had Sloan and Jumla and tiny MCE 08:48.240 --> 08:49.880 and a few others that were gathering 08:49.880 --> 08:52.400 on a regular basis for over a year 08:52.400 --> 08:54.040 to try and think about what we could do 08:54.040 --> 08:57.000 and create a report to provide to the European Union 08:57.000 --> 08:58.360 or to the European Commission. 08:59.440 --> 09:01.200 And we got some interesting ideas. 09:01.200 --> 09:06.040 And I think we actually have the only user research 09:06.040 --> 09:08.640 done for authors on creating better content. 09:08.640 --> 09:12.720 I don't know that anyone else has done user research 09:12.720 --> 09:16.400 on accessibility for authors, which is a real shame, 09:16.400 --> 09:19.760 because this is a real sticking point for problems. 09:19.760 --> 09:22.720 And we don't know why authors are creating terrible, 09:22.720 --> 09:24.120 terribly inaccessible websites. 09:24.120 --> 09:25.720 We're not going to fix the problem. 09:25.720 --> 09:27.960 So we need to have more studies like this. 09:27.960 --> 09:29.840 But that was now over four years ago. 09:29.840 --> 09:31.640 And AI has changed things. 09:31.640 --> 09:33.800 How do we try and think about what is changed 09:33.800 --> 09:35.080 and how we can make this better? 09:35.080 --> 09:45.680 So from the Funker report, there's a few things 09:45.680 --> 09:49.720 that are there that jumped out. 09:49.720 --> 09:52.760 One is the challenges with changing languages. 09:52.760 --> 09:54.080 And that's still a problem. 09:54.080 --> 09:58.360 In fact, Sally, Adam, I've got Adam's last name 09:58.360 --> 10:02.600 as it starts with the C. Sally is a S11Y. 10:02.600 --> 10:06.000 As an open source editing tool, that was the built out 10:06.000 --> 10:10.400 of the University of Toronto, the Metropolitan University 10:10.400 --> 10:12.200 of Toronto. 10:12.200 --> 10:17.880 And he's in working with John from Princeton 10:17.880 --> 10:20.400 looking at creating tools to help identify 10:20.400 --> 10:24.040 when language or parts is being misused. 10:24.040 --> 10:28.200 Because it's all this devicade criteria, 10:28.200 --> 10:32.560 success criteria language is easy to explain. 10:32.560 --> 10:35.160 But there aren't that many language tools out there 10:35.160 --> 10:38.360 that make it easy for people to understand the problems. 10:38.360 --> 10:41.960 But table creators, all text generation, form editors, 10:41.960 --> 10:44.280 they created a list of things that they wanted 10:44.280 --> 10:46.960 to have built into content management systems 10:46.960 --> 10:50.160 and some roadmaps and templates to do that. 10:50.160 --> 10:53.040 And I have an archive of that if people are interested 10:53.040 --> 10:55.800 in seeing the results of that workshop. 10:55.880 --> 11:06.000 So looking ahead, we need to have, I think 11:06.000 --> 11:07.800 that we are going to have to try and find better ways 11:07.800 --> 11:11.600 to use AI to create to help authors. 11:11.600 --> 11:17.600 And it is going to have to go and to be a privacy first approach. 11:17.600 --> 11:21.520 We're going to need to look within the boundaries 11:21.520 --> 11:23.120 of the content management systems that we're using 11:23.120 --> 11:25.400 to see one of the limitations of this. 11:25.400 --> 11:28.600 We're going to have to find ways to improve the quality 11:28.600 --> 11:31.280 and use AI to not just produce more slope. 11:31.280 --> 11:34.040 Because AI can produce content so fast. 11:34.040 --> 11:37.960 But can it actually help authors improve the content 11:37.960 --> 11:39.320 and improve the quality of the content? 11:39.320 --> 11:41.400 Whether it's thinking about different roles, 11:41.400 --> 11:44.120 it might be reading it, whether it's just double checking it 11:44.120 --> 11:47.000 to make sure that you're doing more than what 11:47.000 --> 11:48.280 primarily can do out of the box. 11:48.280 --> 11:52.520 Like there's AI can provide additional information. 11:52.520 --> 11:57.080 And yeah, we want to make sure that when you're using AI 11:57.080 --> 11:58.280 and you're doing your own doc, 11:58.280 --> 12:00.360 you're doing work that information isn't getting leaked 12:00.360 --> 12:06.080 to third party parties, because that can be problematic. 12:06.080 --> 12:09.320 I do think that it's useful to try and think 12:09.320 --> 12:13.040 about having LLMs that where there's a choice of models. 12:13.040 --> 12:14.440 So you're not stuck into one model. 12:14.440 --> 12:16.440 These are changing quite regularly. 12:16.440 --> 12:20.040 And there will be other models that are better. 12:20.040 --> 12:22.760 There, one of the challenges we ran into, 12:22.760 --> 12:26.520 when we were looking at the Funka project with we for authors, 12:26.520 --> 12:29.600 is that you didn't have AI at your fingertips. 12:29.600 --> 12:30.680 It was promised. 12:30.680 --> 12:32.720 It could happen, but it seemed like that was something 12:32.720 --> 12:35.720 that was only going to be available as part of a proprietary 12:35.720 --> 12:36.560 add-on. 12:36.560 --> 12:39.040 We can now build that in with open source tools. 12:39.040 --> 12:43.120 There's a lot of opportunity using small scale, 12:43.120 --> 12:45.520 small language models that can be incorporated 12:45.520 --> 12:48.800 in very specific to our content management systems 12:48.800 --> 12:52.480 to improve the content that we're developing. 12:52.480 --> 12:56.960 So yeah, bundling stuff together, 12:56.960 --> 12:59.600 looking at ways that we can build better prompts 12:59.600 --> 13:02.440 and better systems to try and support more accessibility 13:02.440 --> 13:04.400 out of the box for people who are creating 13:04.400 --> 13:05.200 or editing content. 13:07.720 --> 13:10.720 So a couple of other things in terms of shifting left, 13:10.720 --> 13:13.680 which is a big part of the accessibility movement. 13:13.680 --> 13:16.800 I already mentioned thinking about spell check in Sally. 13:16.800 --> 13:19.000 There's also a tool called editorially. 13:19.000 --> 13:22.280 Editorially is the one that's being built out of Princeton. 13:22.280 --> 13:26.280 This built-in editorially is an integrated into Drupal. 13:26.280 --> 13:28.680 It's sort of into Drupal CMS, not into Drupal Core, 13:28.680 --> 13:31.160 but Drupal CMS. 13:31.160 --> 13:33.840 And these are both tools that are fixing the problems, 13:33.840 --> 13:37.640 many of the problems before they get into the introduction. 13:41.400 --> 13:43.480 And yeah, if you fix the problems earlier, 13:43.480 --> 13:46.560 you're going to save money. 13:46.560 --> 13:50.320 It is way cheaper to fix accessibility problems 13:50.320 --> 13:54.200 by catching them when the author is creating the errors 13:54.200 --> 13:56.160 than it is a few months later, 13:56.160 --> 13:59.320 when somebody comes up in a report. 13:59.320 --> 14:01.680 Those reports are not useful if they're not timely, 14:01.680 --> 14:04.240 because it takes time to go back and fix issues 14:04.240 --> 14:07.080 that were addressed and forgotten about for months ago. 14:08.720 --> 14:11.000 We want to try and find ways to empower users 14:11.000 --> 14:15.200 and see this as being something that isn't another compliance 14:15.200 --> 14:16.880 hurdle that they have to do, 14:16.880 --> 14:18.760 because that's what the policy says. 14:18.760 --> 14:19.960 But something that actually gives them 14:19.960 --> 14:24.280 greater confidence about their ability to deliver content 14:24.280 --> 14:26.680 that is valuable for their users. 14:28.320 --> 14:32.080 We also want to have better ways to have cross-organization 14:32.080 --> 14:32.920 accessibility. 14:32.920 --> 14:35.160 So when you're looking at larger organizations 14:35.160 --> 14:37.160 to make sure that you're using a consistent tone 14:37.160 --> 14:40.480 or a consistent, how do we build the fact 14:40.480 --> 14:43.200 that we have many sites in our organizations? 14:43.200 --> 14:45.200 Sometimes they'll be like universities 14:45.200 --> 14:47.280 or cities of hundreds of websites 14:47.280 --> 14:50.760 that all have to speak to the same audience, the same user. 14:50.760 --> 14:54.840 And we can use AI to try and help build a better user 14:54.840 --> 14:57.880 experience for people, even for jumping across multiple websites. 14:59.400 --> 15:00.400 More is possible. 15:00.400 --> 15:03.720 I mean, all texts is an easy one in many ways. 15:04.720 --> 15:06.760 Two years ago, I gave a presentation of Faustem 15:06.760 --> 15:10.840 about Alt Text and comparing human generated alt text 15:10.840 --> 15:15.560 versus AI-generated alt text, even for government websites. 15:15.560 --> 15:21.160 Most human generated AI is not as good as what I've seen 15:21.160 --> 15:23.840 already, what I saw two years ago with AI. 15:23.840 --> 15:26.680 So that should be an easy one. 15:26.680 --> 15:30.000 Captions and plain language and in simplification 15:30.000 --> 15:31.440 content also useful. 15:33.040 --> 15:35.640 I think we often think about ways of reducing friction 15:35.640 --> 15:39.240 with technology, but I think we also need to think 15:39.240 --> 15:41.800 about how do we increase the friction? 15:41.800 --> 15:44.720 How do we find out how do we find those places 15:44.720 --> 15:46.840 or we actually need to have authors spend the time 15:46.840 --> 15:49.360 to do the thing to verify what they're looking at? 15:49.360 --> 15:51.640 Just like the terms of service agreements 15:51.640 --> 15:54.000 that we all skip through in-depth click on the bottom 15:54.000 --> 15:54.960 and say, yes, we read this. 15:54.960 --> 15:58.240 It's like, no, no, you actually, you have to read this. 15:58.240 --> 15:59.160 You have to do the work. 15:59.160 --> 16:01.920 You don't have to just scroll down and hit submit. 16:01.920 --> 16:04.240 We need people to actually understand 16:04.240 --> 16:07.640 that when you're dealing with AI-generated content 16:07.640 --> 16:11.640 in particular, you can trust, but you have to verify. 16:11.640 --> 16:14.040 Again, that's a Ronald Reagan quote, if you're curious. 16:16.360 --> 16:19.000 So, yeah, and again, how do we provide 16:19.000 --> 16:21.160 these structural support of hands 16:21.160 --> 16:24.520 as opposed to, as people are developing things? 16:24.520 --> 16:29.520 This is definitely a part of a science 16:30.600 --> 16:33.160 but a very, it should be a rigorous process 16:33.160 --> 16:37.360 of looking at how authoring tools are set up, 16:37.360 --> 16:40.720 looking at the, also looking at evaluation tools 16:40.720 --> 16:42.680 for sites that are built. 16:42.680 --> 16:45.000 Are people here familiar with WKGEM, 16:45.000 --> 16:47.560 the WKG evaluation methodology? 16:49.160 --> 16:50.160 Not many. 16:50.160 --> 16:53.920 Again, we should have a lot more people who know what this is. 16:53.920 --> 16:57.560 Hitty certainly knows what this is, but they say, 16:57.560 --> 17:00.880 because you worked out it for many years with WC3. 17:00.880 --> 17:02.920 But this is an evaluation methodology 17:02.920 --> 17:05.760 to try and ensure that sites that are built 17:05.800 --> 17:08.120 are evaluated in the same way, 17:08.120 --> 17:10.000 so that you have a consistent evaluation 17:10.000 --> 17:15.000 with automated manual and human research to try and say, 17:15.000 --> 17:17.520 how should you be evaluating accessibility? 17:17.520 --> 17:19.640 How should your sites be ensuring 17:19.640 --> 17:23.440 that they're structured and consistently evaluated? 17:23.440 --> 17:25.840 And it's a really excellent approach. 17:26.760 --> 17:29.120 I did in the talk and give a slight indication 17:29.120 --> 17:32.400 that I was going to be using AI with WKGEM, 17:32.400 --> 17:34.800 and I had hopes to be able to do that. 17:34.800 --> 17:36.560 I didn't quite get there. 17:36.560 --> 17:42.560 I did get to doing using AI to create an accessibility 17:43.080 --> 17:45.800 conformance report, which is similar, 17:45.800 --> 17:48.000 but it's more towards the USA, V-PAT, 17:48.000 --> 17:49.880 voluntary product accessibility guidelines 17:49.880 --> 17:53.520 for the framework and not in the same way as, 17:53.520 --> 17:55.400 it's not structured in the same sort of format 17:55.400 --> 18:01.640 as the WKGEM, but there are ways that we can start 18:01.640 --> 18:04.560 using AI now to try and gather this information 18:04.560 --> 18:08.120 to provide regular reports and to consolidate the information 18:08.120 --> 18:12.320 so that we have a mechanism to be able to look ahead 18:12.320 --> 18:16.000 in more quickly evaluate progress over time. 18:16.000 --> 18:21.240 Because it's not, there's a phrase and accessibility, 18:21.240 --> 18:24.760 like it's progress over perfection, 18:24.760 --> 18:27.560 but we need to be proving progress over perfection. 18:27.560 --> 18:29.400 We need to do this over and over again 18:29.400 --> 18:31.480 and be able to compare results over time. 18:32.400 --> 18:35.800 And all of these reports are things 18:35.800 --> 18:37.800 that are going to need to be fed back into things 18:37.800 --> 18:39.840 like the Web Accessibility Director 18:39.840 --> 18:42.200 and the European Accessibility Act. 18:42.200 --> 18:43.960 How many people have here run businesses 18:43.960 --> 18:45.200 or a part of businesses? 18:46.280 --> 18:51.280 And do you sell services to the fall into the European citizens? 18:53.400 --> 18:57.760 So underneath the European Accessibility Act, 18:57.760 --> 19:00.720 chapter 11.7, I think, 19:01.560 --> 19:05.400 is chapter five, I think, too many different chapters. 19:05.400 --> 19:08.120 There's a requirement to go off and start documenting 19:08.120 --> 19:10.000 and organizing your accessibility. 19:10.000 --> 19:11.680 What are your performance methods? 19:11.680 --> 19:13.960 What are you doing to confirm that you're applying 19:13.960 --> 19:17.680 to the European Accessibility Act? 19:17.680 --> 19:19.920 We need to start looking at that and making sure 19:19.920 --> 19:22.960 that we actually have documentation 19:22.960 --> 19:25.600 about how we've, about the work we've done 19:25.600 --> 19:28.120 to meet these accessibility requirements 19:28.200 --> 19:29.360 and to comply with them. 19:30.800 --> 19:32.760 So yeah, all these can be useful 19:32.760 --> 19:35.760 to try and help produce reports that are necessary 19:35.760 --> 19:38.120 for compliance, but it should be about providing 19:38.120 --> 19:39.840 better content. 19:41.720 --> 19:43.000 Before I talk a bit about that, 19:43.000 --> 19:44.520 in terms of trying to automate it 19:44.520 --> 19:45.920 and the challenges with this, 19:49.160 --> 19:52.560 it is we wanted to be able to ultimately put things into a 19:52.560 --> 19:56.200 adjacent format or some other machine readable formats 19:56.200 --> 19:59.680 so that it can be easily read and consumed 19:59.680 --> 20:02.920 and compared year over year with machines. 20:02.920 --> 20:05.440 That's something that we need to start moving towards 20:05.440 --> 20:07.040 and then it's great to see the WC3 20:07.040 --> 20:09.440 starting to embrace that type of approach. 20:10.600 --> 20:14.000 We also need to have more testing 20:14.000 --> 20:16.800 and especially testing with users with disabilities. 20:16.800 --> 20:19.680 I do not have a disability, at least not one 20:19.680 --> 20:21.840 that stops me from using the web, 20:21.840 --> 20:25.680 but there is a huge amount of value 20:25.840 --> 20:28.680 from involving people with lived experience of disability 20:28.680 --> 20:31.000 and hiring them to go off and to help ensure 20:31.000 --> 20:33.040 that they can use the products 20:33.040 --> 20:35.560 and that they can overcome all the things 20:35.560 --> 20:37.840 that the automated tools will not see 20:37.840 --> 20:39.720 because the automated tools are great 20:39.720 --> 20:41.800 but they only catch about a third of the issues. 20:45.120 --> 20:48.920 So I do want to challenge us to try and move together on this. 20:48.920 --> 20:50.720 We've got a lot of different people 20:50.720 --> 20:54.880 who know different tools and use them in different ways, 20:54.880 --> 20:58.400 but part of what the WC3 author's cluster was starting to do 20:58.400 --> 21:00.200 was saying, how do we collaborate? 21:01.800 --> 21:04.400 It's something where the offering interface 21:04.400 --> 21:06.040 is always going to be overlooked. 21:06.040 --> 21:08.880 It's not something that generally people are going to pay for 21:08.880 --> 21:12.000 and yet it is a huge part of solving the accessibility issue 21:12.000 --> 21:14.520 because again, it's like spell check. 21:14.520 --> 21:17.600 If we were running spell check like we do accessibility, 21:17.600 --> 21:18.680 it would be nuts. 21:18.680 --> 21:20.440 Like I think people would lose their minds. 21:21.440 --> 21:26.680 And so, yeah, take a look at A-tag part B 21:26.680 --> 21:30.080 and see if there's any you can apply to your CMS. 21:30.080 --> 21:34.360 If you see the, if you're interested in learning more about this, 21:34.360 --> 21:36.920 there's now a community group with the WC3 21:36.920 --> 21:39.520 that you can join if you want to learn more and discuss this. 21:39.520 --> 21:43.920 So I do recommend that you reach out to the WC3 21:43.920 --> 21:47.120 and I've got links to all this in the resources which are next. 21:47.280 --> 21:49.840 But do you take a look at the, 21:49.840 --> 21:51.800 considering the A-tag approach 21:51.800 --> 21:54.800 or so you're considering joining the A-tag community group. 21:54.800 --> 21:56.320 You don't have to be invited to expert 21:56.320 --> 21:57.440 to join that community group 21:57.440 --> 21:59.000 and it's a wonderful place to learn. 22:00.160 --> 22:02.600 I do want to encourage people also to share their work 22:02.600 --> 22:05.680 and to talk with each other and to try finding ways 22:05.680 --> 22:09.520 to go off and improve the accessibility of their interfaces. 22:09.520 --> 22:11.040 Even if you're not a dribble user, 22:11.040 --> 22:13.920 there's a lot of issues that you can look at in dribble 22:13.920 --> 22:15.920 or even issues you can contribute to 22:15.920 --> 22:17.640 but well, how could this be done? 22:17.640 --> 22:18.520 What could we learn? 22:18.520 --> 22:21.840 How could we fix this in whatever scene as you're using? 22:24.560 --> 22:27.680 Also, that's a quick, provide a brief book announcement. 22:28.560 --> 22:33.080 I wrote a chapter in a book on disability 22:33.080 --> 22:38.320 and a digital accessibility ethics with a lady find gold 22:38.320 --> 22:41.960 and a 30 other authors and it's not out yet. 22:41.960 --> 22:44.040 You can pre-order it though. 22:44.040 --> 22:45.600 I think there's links to that as well. 22:46.160 --> 22:50.320 But I wrote a book on disability and open source 22:50.320 --> 22:54.320 and trying to, except I did not include the chapter in this reference. 22:54.320 --> 22:56.400 This is a bad, bad editing job. 22:56.400 --> 22:59.440 But I wrote a book on open source 22:59.440 --> 23:01.920 and accessibility and sustainability 23:01.920 --> 23:04.000 and how these are things that we need to start working on together. 23:06.000 --> 23:08.640 So I have, yeah, resources and link, 23:08.640 --> 23:14.240 but most importantly, I have all the slides are online 23:14.400 --> 23:20.120 and it's ox.ca slash p slash three, actually not four. 23:20.120 --> 23:20.920 That's a lie. 23:23.400 --> 23:25.600 Any other questions? 23:25.600 --> 23:26.320 Excellent. 23:26.320 --> 23:27.520 Thank you. 23:27.520 --> 23:28.640 Thank you. 23:28.640 --> 23:29.640 No problem. 23:29.640 --> 23:32.120 I found it's a super expensive. 23:32.120 --> 23:33.240 I am a bit too bad. 23:33.240 --> 23:34.640 I'm too bad for you. 23:34.640 --> 23:35.920 Perfect. 23:36.920 --> 23:40.160 So, one minor comment and one of the 23:40.160 --> 23:43.080 the minor comment is like, at the rules, 23:43.080 --> 23:47.760 I have not, I do use A, A, A, A, A, A, but I don't know. 23:47.760 --> 23:53.680 I have not convinced that they are really really about to talk 23:53.680 --> 23:57.480 and maybe a lot can be done without them. 23:57.480 --> 24:02.280 So I'm under this, but this that we did. 24:02.280 --> 24:06.640 And the other comment is that, as you can imagine, 24:06.640 --> 24:10.040 I am a net-vocate for accessibility like in real world, 24:10.040 --> 24:11.200 where I'm worked. 24:11.200 --> 24:14.560 And what I noticed from people is something very sad. 24:14.560 --> 24:21.880 I think that they would like to that the care of accessibility, 24:21.880 --> 24:25.360 so that in fact they can forget about it. 24:25.360 --> 24:29.920 And this is, in fact, it took me some time to understand 24:29.920 --> 24:34.960 that how painful this is, because in fact, 24:34.960 --> 24:37.040 I value people taking care of it. 24:37.040 --> 24:42.280 I value people thinking about it. 24:42.280 --> 24:45.000 So yeah, I think my question is, 24:45.000 --> 24:47.640 do you think this kind of truth? 24:47.640 --> 24:52.240 Will more like help people to become more aware? 24:52.240 --> 24:57.040 Or do you think that there is a risk of people just following 24:57.040 --> 25:02.040 the two, but then disengaging, because they will completely rely 25:02.040 --> 25:05.000 on the two for not having to think about it themselves? 25:05.000 --> 25:09.400 I have about an hour of comments that I'd like to respond to on that, 25:09.400 --> 25:11.320 but I will not. 25:11.320 --> 25:17.920 So the question is, is twofold, one question was about AI 25:17.920 --> 25:22.040 and AI trusting in AI to produce accessible content. 25:22.040 --> 25:27.240 And the, is it necessary? 25:27.240 --> 25:30.800 And the second one was looking at checkbox mentality 25:30.800 --> 25:33.600 and how do we get over that idea that people just 25:33.600 --> 25:37.440 want to get this done and don't want to go off and worry about that? 25:37.440 --> 25:40.400 In terms of, is it necessary? 25:40.400 --> 25:43.160 I think it probably is. 25:43.160 --> 25:46.120 I didn't think it used to be, but I think that the rate 25:46.120 --> 25:49.360 of change for accessibility, even after legislation, 25:49.360 --> 25:52.320 is still so incredibly painfully slow. 25:52.320 --> 25:55.080 And if we want to see progress, 25:55.080 --> 25:59.200 then I think that AI in specific use cases can be good. 25:59.200 --> 26:00.920 And it's not that AI is perfect. 26:00.920 --> 26:04.240 The question is, is it better than humans? 26:04.240 --> 26:06.600 And not the best humans, because it will never be better 26:06.600 --> 26:08.840 than the best humans I don't think. 26:08.840 --> 26:11.680 But it will be better than the average developer. 26:11.680 --> 26:13.640 And I think in many, certainly with all text, 26:13.640 --> 26:15.720 it's already better than the average developer. 26:15.720 --> 26:20.720 Or the average content author at right now. 26:20.720 --> 26:23.240 So for some things like that, I mean, absolutely, 26:23.240 --> 26:24.680 I think it's necessary. 26:24.680 --> 26:28.480 In terms of the checklist, I think it's a cultural issue. 26:28.480 --> 26:30.920 I mean, accessibility, champion, as well. 26:30.920 --> 26:32.520 In our organization, in other places, 26:32.520 --> 26:34.400 I try and engage with people, 26:34.400 --> 26:36.080 to try and change the mindset. 26:36.080 --> 26:38.080 There's a real movement of people trying to be part 26:38.080 --> 26:39.840 of building champions, movements, 26:39.840 --> 26:43.760 so that it changes the culture from a checklist culture 26:43.760 --> 26:46.200 to a, hey, how can we make this better? 26:46.200 --> 26:49.600 How can we be excited about making our website, 26:49.600 --> 26:54.160 our tools, our community, our staff, more accessible today 26:54.160 --> 26:55.160 than they were tomorrow? 26:55.160 --> 26:57.960 That we could be a better, more inclusive, more diverse, 26:57.960 --> 27:01.160 more productive team, if we think about accessibility. 27:01.160 --> 27:03.320 Even if it's just thinking about our future selves, 27:03.320 --> 27:05.560 that this is something that will benefit all of us 27:05.560 --> 27:09.080 if we make things, if we invest in accessibility. 27:09.080 --> 27:11.240 But that's it, all right? 27:11.240 --> 27:12.240 Thank you. 27:12.240 --> 27:13.240 Thank you, bye. 27:13.240 --> 27:16.040 Thank you, thank you, thank you. 27:16.040 --> 27:18.080 I really, thank you, thank you.