UF: presupuesto y estrategia para video accesible ADA (Full Transcript)

Cómo UF auditó su biblioteca, proyectó minutos, y aplicó una estrategia por niveles (ASR + edición humana) para cumplir ADA Title II.
Download Transcript (DOCX)
Speakers
add Add new speaker

[00:00:00] Speaker 1: Thank you for joining today's session, Building a Budget for Video Compliance with the University of Florida. My name is Eric Ducker. You might have seen me from other webinars, especially part of this series of Countdown to Compliance. I use he, him pronouns and I will be taking a different role today, which is moderating today's webinar, which is new for me. So bear with me if I make any mistakes. With that taken care of, I'd like to welcome today's speaker, Brian Smith. He comes from University of Florida, longtime partner with 3Play Media, and we're excited to have you here with us today. So before we dive in about the topic at hand, let's set some context. Can you share a bit about your role at UF and how video accessibility fits into your team's responsibility?

[00:00:53] Speaker 2: Sure. Thank you, Eric, and thank you 3Play Media for having me here. Like Eric said, I'm Brian Smith with the University of Florida. My pronouns are he, him. I have been the manager for Video and Playvision Services with UFIT for 10 years now. I have worked for the team for going on 25 years, starting as a student. So I have seen many evolutions of video, online streaming, management, accessibility concerns, and all it entails. My group specifically oversees the Mediasite Video Management Platform at UF. We conduct live streaming and hybrid events across campus. And as far as accessibility, that became a larger role for our team a couple years ago, and we support the captioning of videos within the Mediasite platform. We partner with other faculty and departments across campus to make sure that all video content within Mediasite meets ADA Title II compliance.

[00:02:01] Speaker 1: Okay. So that sounds like a lot of UF video. Can you walk us through, maybe just a step back, of what is the university's overall strategy for budgeting and allocating budget across the various campuses that UF oversees? Is it centralized, decentralized, or a mix of both?

[00:02:23] Speaker 2: Sure. UF uses a hybrid model. We provide centralized caption solutions for all content maintained within Mediasite. So the central UF IT budget, UF budget, covers content in Mediasite. Some individual colleges and units still maintain their own video accessibility budgets, either within Mediasite or using other platforms. Within Mediasite, they may have their own decentralized budget if they want to have faster turnaround times or higher accuracy, always human-edited content and stuff like that. So yeah, it's definitely a hybrid between centralized and decentralized at UF.

[00:03:05] Speaker 1: Okay. I'll put a pin in my budget questions for a second around the exchange of cash per se. So, obviously we're here today to talk a little bit about the budget in context of ADA Title II compliance. Can you walk us through, we started talking almost a year ago around this challenge that you guys had. When you first started, what were some of the biggest questions or uncertainties that you had around budgeting for this opportunity, for this challenge?

[00:03:42] Speaker 2: Yeah, well, early on, the biggest unknown we had was the volume of content that we had to caption, how much legacy video content existed and how much new content would be produced each year. At this time last year, when we were making our evaluations, we had over 11 million minutes of content stored on Mediasite. And we knew we had to eliminate or archive millions of those minutes before we could even determine what our budget needs were at that point. So, that was the biggest thing we did early on was evaluate how much content we had and predict how much we had to produce in the future.

[00:04:28] Speaker 1: Okay. So, that's really helpful of like the volume of kind of the existing. What about, how did you kind of prioritize what needed to be archived? What were you focused on?

[00:04:44] Speaker 2: Well, we began with an entire platform level audit of Mediasite using the metadata that the platform provides, like viewership, view history, and age of videos. We already had an established content retention policy, but it had only loosely been enforced. So, any content that was older than two years that hadn't been viewed in the past year was eligible to be archived and under our content retention policy should have been archived. But there was always various hurdles to accomplishing that. For instance, various colleges and units had their own retention needs that didn't fit within the tight scope that we had. Like the College of Medicine, for instance, they may need to maintain videos for at least four years for the entire length of a certain cohort. So, that's really where we've prioritized our, you know, the beginning of this process. Go ahead.

[00:05:46] Speaker 1: Yeah, yeah, no, that makes sense. And did you, you know, obviously the backlog was very large and there's your pre-established retention policies. How is that, has that transformed, you know, nowadays, now that you're producing all this new content, do any of those things change or, you know, how are you kind of viewing the difference between like what you guys do at the back, you know, what you're doing with backlog and what you're doing with kind of go-forward captions, you know, tomorrow's lecture, what's happening tomorrow versus what happened two years ago?

[00:06:19] Speaker 2: Right, so anything that's produced from here on forward is going to have captions. So, that's just part of the strategy. Yeah, any new content is going to get captioned. So, it really came down to, you know, getting us to a point where we had only the content online that is new enough or will be used in the future.

[00:06:48] Speaker 1: Okay, that makes sense. So, kind of going on off that with the future video production, a little bit more ambiguous in terms of how much you're going to have, what were some of the tactics that you used to predict how much content you're going to be looking at in the future?

[00:07:06] Speaker 2: We reviewed the historical data over just the past three years, you know, going too far back beyond that, we started to hit the COVID skewed numbers. And we actually saw a rather consistent amount of content being produced each year, ranging from about 1.2 to 1.4 million minutes of content, new content each year. So, based on that, we were able to, you know, fairly, hopefully accurately estimate, you know, the future content being produced this year with a little bit of overhead added on.

[00:07:41] Speaker 1: Yeah, did you, over the last three years, I'm just curious, did you guys see a trend line of any growth in content or was it really at this point you can't tell?

[00:07:51] Speaker 2: It was a very shallow increase.

[00:07:53] Speaker 1: Yes. Okay.

[00:07:56] Speaker 2: But yeah, I mean, it was rather consistent, but there was a slight increase year over the year.

[00:08:01] Speaker 1: Okay. And then thinking about the, I'm sorry. When you were kind of doing this audit of captions, or sorry, of video content, did you uncover like any hidden craters? I know this is kind of centralized around media sites specifically, but did you kind of uncover any like, oh, I didn't know that these people were using media site? Like what were some of the potential surprises you found?

[00:08:38] Speaker 2: Fortunately, I've been very hands-on in depth with the media site platform for a long time. So I'm fairly well in tune to, you know, what content creators are on there. You know, we have, you know, discovered content creators that aren't using media site that are uploading videos directly to campus, or, you know, they may have a small instance of Kaltura and other video platforms out there. So not necessarily hidden creators within media site, but definitely, you know, pockets of creators in various other places on campus that aren't using media site. And we've done our best to at least educate those other users on, you know, the regulations that are forthcoming.

[00:09:23] Speaker 1: Yeah. Yeah. So that's a good quick question. And then just a side question of, you know, with this kind of push and centralization of budget, did you try to push for more people to kind of come within, like join the media site ecosystem that were not in it? Or, you know, was there any kind of conversations of, I guess, increasing the pie of responsibility? I know that's probably not what you really wanted to do, but I'm curious if that type of conversation happened.

[00:09:52] Speaker 2: Our focus primarily right now has been to, you know, encapsulate what we know we're going to have. But, you know, we have seen an increase as this semester has gone on with people, more and more people requesting access to use media site. So just naturally based on, you know, communication and education, we have seen more users come onto media site from other platforms. Yeah. Okay.

[00:10:16] Speaker 1: That's good. And then not to, you obviously have been very clear media site. Do you know, like how is the university managing their YouTube content? How is the university managing like the non-media site content? Is it truly department specific, or is there other kind of entities that you've been kind of advising on best practices there as well?

[00:10:41] Speaker 2: It's very department, you know, dependent on if you're using other platforms, or, you know, it may not even be department specific. It may just be a single faculty member choosing to go off on their own and using another platform like YouTube. But we have, you know, various means to communicate with them. We have a great team on campus, the accessibility community of practice that, you know, meets monthly and with different leaders across campus to help get the word out about the accessibility needs. There are various faculty newsletters that have been touching on the subject for the past few months. So what just, the campus as a whole is doing its best to educate all, you know, content creators out there.

[00:11:29] Speaker 1: Yeah, okay. So kind of shifting gears a little bit to kind of the operational aspect of kind of organizing this budget. Obviously, like your team centered around the kind of source of truth of what was required or what was needed. Who else though was like involved in evaluating and planning for these budgets across, you know, whether it's your team or other teams that you had to interface with?

[00:11:58] Speaker 2: Primarily the, you know, the planning has been done by my team itself, but we absolutely have involved others such as our Electronic Information Technology and Communication Accessibility Officer, EITCA. It's a mouthful, sorry. And the leadership within UFIT itself to determine, you know, the budget for this current year. You know, we evaluated various solutions within our own department with input from various content creators and the colleges and, you know, instructional designers from UFIT's Center for Instructional Teaching and Technology. So we've definitely got input on the tools and how to best go about it, but the budget recommendations came from my team itself.

[00:12:44] Speaker 1: Okay, and did you have to interface at all with, you know, the legal teams or kind of other ancillary administrative roles that, you know, oversaw kind of compliance in general at the university level?

[00:13:00] Speaker 2: Yeah, our EITCA officer does most of the go between between us and, you know, say the general counsel's office or the compliance office. So she was our main point of contact as far as any, you know, legal regulations or, you know, evaluation of, you know, what the guidelines were. But, you know, we did very in-depth studies on WCAG itself. So, you know, our department had, you know, a couple at this point, subject matter experts when it comes to WCAG. So we pretty had a straightforward idea when we got down to it of that, if the content was on media site, if it was viewable, it had to be accessible. Okay. So there's no question about that in our team.

[00:13:46] Speaker 1: Yeah, that's interesting. Okay. So I know that today you haven't budgeted explicitly for audio description in the same way that you have for closed captioning. Was there any discussion this year, this budgeting cycle around audio description? And maybe I'll come back to that later in the discussion, but, you know, just curious from these conversations with legal and compliance, like what were some of the nuggets around the audio description and where you guys are, where were you guys in that budgeting cycle for it?

[00:14:19] Speaker 2: In agreement with our directors, you know, we decided as far as the audio description goes to focus on educating and encouraging faculty to actually do audio description within their teaching. So if they're presenting slides, you know, describe what's on the slides. So any audio description, we're not focusing on adding it to the videos, but incorporating it as it's being created.

[00:14:48] Speaker 1: Yeah, no, that makes sense. And that's a very common strategy. And when we talk about like where you're headed, I'm gonna bring this back up. So I'm gonna put you a little bit on the spot. So thinking about like how you got to that yes, like what were kind of the key partners in that process for you, where you were kind of building the momentum to get to a yes on your budget for this fiscal year, you know, going into the compliance year of 2026?

[00:15:23] Speaker 2: For the most part, we, you know, try to just make it as seamless as possible for our constituents. That's why we sought automated solutions. We got input from various different colleges that have their own, you know, content creators that, you know, manage content within Mediasite and got their input on that. They were our main partners, you know, the small handful of, I like to call them power users of Mediasite, along with our instructional designers in the Center for Instructional Teaching Technology. They were our main partners in all the decision-making.

[00:16:06] Speaker 1: Okay, and then ultimately, like who at the end of the day signed the piece of paper that said, all right, Brian, your team gets this allocation of money for, I'm sure it was probably a broader budget, but like the carve out of, you know, the video accessibility needs that your team needed.

[00:16:26] Speaker 2: We were actually very fortunate that we had an existing budget for captioning needs. Like you mentioned at the beginning, we've had a partnership with 3Play for a while. So that budget was mostly used to accommodate, support accommodation requests, but also content created for courses that were 80% online and would be reused in future semesters. So that money had a specific captioning purpose. But with the help of our vendors, we were able to find a solution that would fit into that pre-existing budget and allow us to caption all of our existing content and content we anticipated to be created this year. And once those vendor solutions fell into place, you know, UFIT leadership, you know, at the very top with our CIO, they readily agreed to approve the same funds that we had last year to meet our needs this year.

[00:17:21] Speaker 1: Okay, so I'm kind of curious, like, you know, it's December now, I'm sure you're looking at next year already a little bit. What are you kind of gearing up for going into the next budget cycle? Are you looking to see how you can do more with audio description? Are you looking to do more coverage of captioning, you know, beyond ASR in certain cases? Like, what is it that you're kind of looking to gear up for now that you kind of got the solution in place today?

[00:17:56] Speaker 2: Yeah, we haven't really touched on it too much yet, but, you know, we're making heavy use of 3Play's predicted caption accuracy technology, which allows us to automatically, you know, ASR-created content up to a human-edited level. And, you know, we've worked closely with 3Play over the past few months to, you know, find the right balance of when that changeover happens. So we're hoping that once we get our backlog of content captioned by April, we can focus on maybe improving that threshold and maybe sending more off to human-reviewed captioning just so we can increase our accuracy level across the board. Okay. So that's going to be probably the main focus for the next year of planning is, you know, where is that balance between accuracy and cost?

[00:19:02] Speaker 1: Yeah. So I think that brings me to the next kind of section here of like, and you kind of hinted at the solution, but I'm curious kind of like, what was part of like the evaluation process, you know, of the different captioning tools or vendors? Like how, you know, a million plus minutes was, you know, 10X what you were doing today, you know, in the current state, like how did you balance kind of the price and the reality of your budget in front of you?

[00:19:33] Speaker 2: Yeah. Well, we knew that no single tool would meet every need. So our evaluation focused on accuracy, turnaround time, integration, scalability, support responsiveness, and cost. You know, all those factors came into play. We ran pilot evaluations with, you know, over half a dozen different vendors, you know, sending the same videos across to each of them so that we could, you know, compare the actual outputs. You know, we knew that balancing price with quality required a tiered strategy rather than a one-size-fits-all approach. So, you know, through a combination of automated captioning, QMED edited captions with the help of that predicted caption accuracy tool from 3Play and the captioning solution provided within MediaSite itself, we were able to find that balance.

[00:20:26] Speaker 1: Okay. So just for the audience here, the predicted caption accuracy tool basically allows you to measure at the file level what the expected accuracy of the ASR output is. That allows you to dynamically, you know, automatically in this case, make a decision point of, should I upgrade it because it falls below my recommended threshold of accuracy? Or should I keep it because it's above that threshold? How did that conversation kind of go down in terms of like justifying that some of your captions will not be compliant, but they will be of certain accuracy minimums at least? What were some of those conversations internally like?

[00:21:12] Speaker 2: Yeah, we knew we had to find a balance, you know, it would be great for us to have 99% accuracy across the board, but with, you know, in the end about 6.5 million minutes that we had to caption within this calendar year, we knew we had to have a little bit of sacrifice and just, you know, do our due diligence and show, you know, best effort to meeting the letter of the law. Don't quote me on that. I know this is transcripted and all that, but so what it really came down to was, you know, kind of almost trial and error to find, you know, the right balance. We went through a few months of that process with 3Play. We tweaked to find that right level, but for any accommodation request content, we were sending that still to human reviewed off the bat. We weren't making use of any predictive accuracy scores. So, you know, that content we were prioritizing. Yeah. Yeah.

[00:22:25] Speaker 1: Is there, so that's a good question. So obviously student accommodations like a hand raiser that gets sent directly, that makes a lot of sense. And that's what we see pretty much across the board, across the university landscape. Are there any other exceptions that you guys have looked at in terms of, it's not necessarily about the quality of the ASR, it's about the importance of this piece of content having really, really compliant and accurate captions?

[00:22:48] Speaker 2: Yeah. So one case is any live event we do, we use 3Play for live captioning. So that's our automatically already getting, you know, a higher level of service. And then a lot of the training videos that we create, the one-off messages from the president, a lot of the stuff that is highly visible is being prioritized with, you know, straight to human captions. So we are finding that balance with a few different sources of content.

[00:23:26] Speaker 1: Okay. So I guess one of the questions I always have is on the faculty side, obviously that's a really important, you know, partner in this whole ecosystem. What's been kind of the message to faculty with this solution in place? Did you, what information are you providing? What is it, what's that relationship, you know, now versus what you had maybe, you know, in the spring when you didn't have this solution in place? What were some of the changes that you've had with kind of your faculty partners?

[00:24:02] Speaker 2: Well, communication has been the biggest difference as far as, you know, getting the word out about accessibility. And it's not just about, you know, video accessibility necessarily. You know, WCAG includes all forms of digital media that has to be accessible. So it's been a broader message going out to faculty and other, you know, constituents on campus. And, you know, as I mentioned before, we have the accessibility community practice working group. We have, you know, various faculty newsletters and we have multiple different departments with webpages. And we've worked with them to make sure that information about accessibility is readily available. Okay.

[00:24:48] Speaker 1: That makes sense. Okay. So a couple more questions. I do want to kind of talk a little about audio description if you don't mind. But so maybe we just start there a little bit. I'm just seeing a few attendee questions and I think audio description is still a lot of on people's minds. I think the first question is like, I guess the open question is, what are you guys doing about the backlog for audio description? So like, obviously you can't change what the university or the professor already taught and how they taught it. Is there any priority decision around how to make sure that certain pieces of content are adequately accessible or from an audio description perspective, or what is kind of the university stance today? And, you know, is there some ground roots activity ground roots activity to kind of help figure out how to find funding for that more long-term?

[00:25:50] Speaker 2: Fortunately, I do not have a good answer for that because we are not doing anything with the backlog as far as audio description goes. It is possible to put audio descriptions to media site videos, but we just, we haven't gone down that road yet. It's been such a daunting task just to get everything captioned. So I'm sure those conversations will come up after the first of the year, but we just haven't got there yet.

[00:26:21] Speaker 1: Yeah, okay. And then there's one other kind of interesting question that's not directly related to audio description, but more of, you know, Florida being home to a pretty bilingual population. I'm curious, is there been any push in just making more Spanish content available or localizing more of the English first content into Spanish for the campus?

[00:26:48] Speaker 2: Yeah, it hasn't necessarily been a huge issue at UF. That probably would be a bigger issue with some of the Southern Florida schools. But with the media site ASR solution, it does auto detect different languages. So if a course is taught in Spanish or a course is taught in Polish, that tool has been great for us as far as captioning content in native languages.

[00:27:19] Speaker 1: Okay, okay, yeah, that makes sense. This is another good question that kind of popped up around the budgeting and getting, are you aware or is there anything from like the government level of Florida that has like looked at how do you find grants to fund some of these activities? You know, there's the federal law and then there's the, how do we actually get it executed and on the ground level? Has there been any clear path to like from the, you know, have you, there's been any indication from, you know, the Florida government to support some of these actions or is this entirely on UF and its budget to figure out?

[00:28:01] Speaker 2: It's, at this point, it's been entirely on UF and its budget. As far as I know, there hasn't been any talk of grants or anything else that could help us with our budget. Yeah, so.

[00:28:16] Speaker 1: Okay, what about, so I'm gonna, there's a good question here around, obviously like there's lecture content and kind of the need for audio description. Do you guys make, you know, when you get videos from third parties, do you have any kind of stipulations or requirements around making sure that you get audio description as kind of part of that bundle so that you're at least getting some of the third party content with proper audio description?

[00:28:47] Speaker 2: It really hasn't been a focus for us at all. You know, other than getting the communication out to the faculty that they, you know, should be doing audio description within their newly created content. It just really hasn't been a point of emphasis for us.

[00:29:03] Speaker 1: Okay, yeah, and then I think there's an interest from our audience if like, if there's a link or a material of like, how are you are communicating or what material you are using to kind of train faculty on audio description and, you know, best practices around universal design. I think that would be super helpful. And we can share that out. If that is publicly available, we can share that out with the audience post webinar.

[00:29:30] Speaker 2: Yeah, I mean, a good starting point would be just simply our accessibility website. It's accessibility.ufl.edu. So that is a publicly accessible site. So that would be a great starting point. It doesn't focus only on video or only on, you know, digital media, but it's a good starting point. And then there are resources on our website, which is it.ufl.edu slash video. I'll put, I'll type out those links in here in a minute, but that also has resources on accessibility.

[00:30:11] Speaker 1: Okay, a few more questions for you. You know, on a percentage basis, are you comfortable sharing just on a percentage basis of like how much is like video accessibility, like part of your like overall budget in like maybe the IT group?

[00:30:37] Speaker 2: It's probably very minuscule in the percentage of, you know, the entire IT budget. Yeah. But it's a significant portion of our budget. I mean, for my department, it probably makes up 90% of, you know, what we spend in the year.

[00:30:56] Speaker 1: Okay, okay, yeah. So in the grand scheme of things, pretty minuscule, but at the micro level of like solving accessibility, it'd be a pretty big, pretty big expense overall. So definitely fair. And then there was one other question that was interesting from the attendees I wanted to highlight. So do you, what hap, like, I guess, I mean, at 1.3, 1.4 million minutes, an unexpected spike would be pretty significant amount of content. Is that, you mentioned you have this over, a little bit of overhead air cover. Is that kind of the intention is to cover any kind of unexpected, hey, content, someone just uploaded a ton of content unexpectedly this semester?

[00:31:58] Speaker 2: Yeah. We, like I said, we have a little overhead, you know, probably an extra, you know, 100,000, 200,000 minutes of overhead. But, you know, as we continue to fine tune that, you know, threshold of the predicted accuracy score, you know, that those minutes may or may not be used due to that. So it is a fine line that we're going to be towing. And, you know, we may have to send any unexpected content more towards the media site captioning ASR, because that's a more reasonable price point for bulk content, but for the newly created content, you know, we're definitely trying to use the three play avenue.

[00:32:54] Speaker 1: Yeah. So with that in mind, kind of going into the next budgeting cycle, you obviously didn't really have this level of data prior. You had some theories and stuff. How much do you think you can use this next, this year's data to kind of influence, hey, if I get X amount more money, we can have, you know, this many more files protected against non-compliance, you know, captions. Like how, what level of granularity do you think is going to be valuable to position in front of your finance teams?

[00:33:31] Speaker 2: Oh, I mean, it's amazing that, you know, you use the term granularity that we have on both the media site and the three play side. We've always got great analytics from media site, but we are making heavy use of the reports provided from three play at this point that, you know, shows us what the accuracy is per video, how many of these videos are being sent off for human edited. So, I mean, we just have so much information at our fingertips now that would definitely inform our future decisions and budget requests.

[00:34:11] Speaker 1: Yeah, that makes sense. So I kind of want to leave it off with this one kind of future looking, or I guess advice question. Obviously UF is one of the larger institutions in the United States for education. And honestly, historically has been a pretty strong advocate of accessibility. They have a really strong disability resource center. I've personally visited, it's a really great service and there's a lot of care in digital accessibility in the Florida system, UF Gainesville campus. What advice would you give other institutions, you know, small or large trying to build kind of that realistic and defensible budget as they come up against this deadline and kind of making sure that this is incorporated into their go forward budgets?

[00:35:09] Speaker 2: You know, as we just touched on, most important advice I can give is, build your budget on evidence, not assumptions. The compliance is as much about demonstrating due diligence and institutional intent as the amount of dollars you spend, but make those decisions based on whatever information you have available. That's the best advice I can give.

[00:35:36] Speaker 1: Yeah, I think that's so important. Like, you know, we all are in our own little niche world all of the time, and then oftentimes we forget we are just a part of a much bigger pie and we have to be super, super, like in my own job, I have to be extremely on point with data to get someone who's looking at my work 5% of the time of their job to really understand. And I think that's an important piece, Brian, that you just brought up, which is accessibility is every day for us, but it's not every day for the people that are writing the checks at the end of the day for everyone.

[00:36:16] Speaker 2: And I'll repeat a note that just came across the chat. It happens to be from my ETICA officer. Compliance is cheaper than lawsuits and loss of reputation. So thank you, Anne, for that one.

[00:36:28] Speaker 1: Love that. So thank you so much, Brian. I think we had a great conversation, great information, great information. Thank you everyone for asking great questions. I hope I try to incorporate as many as I could into the flow of the conversation. So thanks again, and I hope everyone.

ai AI Insights
Arow Summary
El webinar trata sobre cómo la Universidad de Florida (UF) está construyendo y justificando un presupuesto para cumplir con ADA Title II en accesibilidad de video, centrado en el captioning dentro de Mediasite. UF usa un modelo híbrido: un presupuesto central cubre el contenido en Mediasite, mientras que algunas unidades mantienen presupuestos descentralizados para mayores niveles de servicio (más rapidez o captions siempre con edición humana). El mayor reto inicial fue cuantificar el volumen: había más de 11 millones de minutos de video almacenados; primero tuvieron que auditar y aplicar una política de retención para archivar contenido antiguo y poco visto (p. ej., >2 años y sin vistas en el último año), con excepciones por necesidades académicas (como Medicina). Para proyectar el gasto futuro, analizaron la producción de los últimos tres años (evitando sesgos de COVID) y estimaron 1.2–1.4 millones de minutos nuevos/año con un pequeño margen.

En la evaluación de proveedores y herramientas, UF consideró precisión, tiempos de entrega, integración, escalabilidad, soporte y costo, concluyendo que se requiere una estrategia por niveles. Usan una combinación de ASR y captions con edición humana, apoyándose en la tecnología de “predicted caption accuracy” para decidir automáticamente cuándo escalar de ASR a edición humana. El contenido por solicitudes de acomodación, eventos en vivo (live captioning) y piezas de alta visibilidad (mensajes presidenciales, trainings) se prioriza directamente a servicio humano. Para audio descripción, este ciclo no presupuestaron un backlog; el enfoque actual es educar a docentes para incorporar descripción durante la creación (describir diapositivas/elementos visuales). Reconocen que la conversación sobre audio descripción probablemente aumentará después.

Operativamente, el plan y recomendación presupuestaria provienen del equipo de Video/Playvision Services, con participación del oficial de accesibilidad EITCA, liderazgo de UFIT, diseñadores instruccionales y “power users” de Mediasite. La aprobación fue facilitada por contar con un presupuesto histórico de captions (antes orientado a acomodaciones y cursos reutilizables) y por encontrar una solución que encajara en ese monto; el CIO aprobó mantener fondos. Para el próximo ciclo, planean ajustar el umbral de precisión y posiblemente enviar más contenido a revisión humana una vez cerrado el backlog. La recomendación clave para otras instituciones: presupuestar con evidencia y datos, no con suposiciones; además, el cumplimiento es más barato que litigios y daño reputacional.
Arow Title
Cómo la Universidad de Florida presupuestó el cumplimiento ADA Title II para video
Arow Keywords
Universidad de Florida Remove
UFIT Remove
Mediasite Remove
accesibilidad digital Remove
ADA Title II Remove
WCAG Remove
captioning Remove
subtítulos Remove
ASR Remove
predicted caption accuracy Remove
edición humana Remove
presupuesto híbrido Remove
retención de contenido Remove
auditoría de video Remove
live captioning Remove
audio descripción Remove
cumplimiento Remove
gestión de riesgo Remove
Arow Key Takeaways
  • UF usa un modelo híbrido: presupuesto central para Mediasite y presupuestos locales para necesidades especiales.
  • El primer paso para presupuestar fue auditar el inventario (11M+ minutos) y reducirlo con una política de retención basada en edad y uso.
  • La proyección anual se basó en datos de 3 años (1.2–1.4M minutos/año) evitando distorsiones por COVID, con margen adicional.
  • La estrategia efectiva es por niveles: ASR + escalamiento a edición humana según umbral de precisión y prioridades.
  • La tecnología de predicción de precisión permite decidir automáticamente qué videos requieren upgrade a edición humana.
  • Acomodaciones, eventos en vivo y contenido altamente visible se envían directamente a captioning humano.
  • Audio descripción aún no se aborda para backlog; el enfoque actual es formación docente para describir contenido durante la grabación.
  • La toma de decisiones involucró a EITCA, liderazgo de UFIT, diseñadores instruccionales y usuarios avanzados; el equipo de video lideró el presupuesto.
  • Mantener un presupuesto histórico de captions facilitó la aprobación ejecutiva al demostrar continuidad y viabilidad.
  • Consejo clave: construir el presupuesto con evidencia (analítica y reportes) y recordar que cumplir suele costar menos que litigios y daño reputacional.
Arow Sentiments
Neutral: Tono informativo y pragmático: se discuten retos (gran volumen, costos, backlog) y soluciones basadas en datos, con énfasis en gestión de riesgo y cumplimiento, sin carga emocional marcada.
Arow Enter your query
{{ secondsToHumanTime(time) }}
Back
Forward
{{ Math.round(speed * 100) / 100 }}x
{{ secondsToHumanTime(duration) }}
close
New speaker
Add speaker
close
Edit speaker
Save changes
close
Share Transcript