Emerging Output Technologies

Hello all, welcome back to our course on digital accessibility. And uh in the previous session we spoke about emerging input technologies and what are the limitations of the traditionally known input technologies. And in this session we will talk about emerging output technologies and what are the limitations of current or traditionally known uh output technologies. And of course in this session also there will be uh some discussion regarding use of AI and how AI's role is going to impact development of emerging technologies uh because that is largely impacting most of the interaction devices both input and output. Uh so let us dive into it. So I'm sure you are aware of the commonly known output devices. So output devices are meant to give you u perceptible information and um you basically utilize that information. You per you are able to perceive the uh information projected by a certain output device using uh one or many of your modalities or senses as we call it right. So monitor I'm sure uh which is like a screen is something which is commonly known. Printer can be one which is able to produce like a hard print hard copy print version of digital um media. Speaker can produce audio. Headphones also produce audio. Then projector is also a good media. It's also uh able to um you know kind of project it for a mass audience uh while the screen or uh a printed version may be limited to uh single person use. So before we dive into the limitations of um the the output devices and um the input devices as well, uh this is a video link which I will not play here uh because it contains sensitive uh visuals. You can watch it at your own convenience. And this is a story of Jared who is u um you know suffering with cerebel pulsy and he is still a UX designer um visual designer and what are the various input and output devices how he interacts with computer in his day-to-day life is something which is very inspiring and also very informational uh to watch. Please watch this video at your own convenience. So let us now talk about uh the limitations of traditionally known output devices. So uh they do present significant accessibility limitations because they often rely on a single sensory modality um output. Right? So for example, a screen is only catering to um the visual senses, right? which is like a screen. Um there may be some plugins etc available uh which may enhance the usage uh or assisted technologies available which may make the screen accessible say for a person who's visually impaired uh but for example a screen reader um but the technology itself is not multimodal technology itself is single s single sensory modality oriented it. Same is the case with any speaker output uh which is like a headphone or a speaker like these limitations surely create barriers for individuals with sensory limitations or cognitive impairments and here um in essence like traditional technologies are often designed for an average user which is where I think we've talked about this aspect uh a couple of times in the course or rather aabled user which is what we called ability bias and for that you can go back to uh week one or two where uh details about ability bias uh is discussed and these biases have resulted in the currently used uh technologies. assuming that something which is visually um uh you know communicating and information um is catering to large number of audiences while slowly people are realizing that multimodality is the way to go. Um and how similar in the similar manner how input devices were catering to the operability aspect of the POR accessibility principles. What do you think output devices are catering? Can you think about it? Yes, it is perceivability aspects of the PUR principles come under the output devices. So even in the perceivability uh principle we always talk about multimodal ways of um um presenting information or communicating information to your potential user can make the information perceivable to them in multiple formats and they can choose according to their sensory abilities. So most of the time these traditional um technologies are designed for the aabled user and they fail to be perceivable enough uh to be interpreted reliably by a wider range of users and assisted technologies etc plugins are required to make them accessible. So this is an example of such a technology which is called um enlarging screens. So here there's a book and uh it's like a magnifying glass basically and uh it shows the text in a very big format. So it is primarily for low vision readers. requiring ongoing manual effort. They may be expensive uh to retrofit and even to achieve minimal accessibility standards it may be very difficult with the traditionally known technologies. So let us uh talk a little bit more about what are the issues with the state-ofthe-art uh currently known output tech from different perspectives. So um from a visual impairment perspective uh there there can be lack of alternative text or all text which kind of makes it unavailable for uh even screen reader users to understand what the image is about. Uh so traditional visual displays or screen static images etc do not inherently provide text descriptions for non-ext content. Non-ext content means videos or images making them inaccessible for users relying on screen readers or even braille displays for that matter. Color dependency may be one of the aspects that information conveyed slow solely through color without any labeling uh is also inaccessible for individuals not only with blindness but also with color blindness poor contrast or text sizing. So font fixed font sizes or where they are not uh the users do not have the liberty to uh enhance enlarge the size or minimize the size or inadequate color contrast ratios or lack of text sizing resizing options make the text difficult or impossible to read for people with low vision or color blindness or light perception. Light perception is also a condition which is come which comes under visual impairment. Uh so uh that also where the contrast can be changed in order to make the text a little bit more visible. So these aspects have uh been comprehensively covered under the W CAG because they primarily come under screen paste output. Right? So all of these uh aspects you can again go back to our session on WAG. There are multiple sessions on WAG guidelines. There are several uh sessions on the visual design aspect uh of accessibility. You can go back uh to those sessions to learn a little bit more. Here we are talking about output technologies and their limitations uh from the perspective of different kinds of users and different kinds of needs in case of auditory impairments. uh so audio only content so traditional audio outputs say for example phone systems or public address information systems or announcement systems are only loudspeaker based videos many a times don't have captions uh so they these kind of content or media excludes deaf or hard of hearing individuals as there are no transcripts or captions alternatives provided provided by default. Volume control is also something which is uh very important particularly if the videos or audios are embedded in a website or in an application. There is not a lot of um interface available to increase or decrease the volume. If uh the app or the website enables usage of volume control through the device itself then maybe some possibility but the code itself has to allow that and uh other than that uh there should be um muting mechanisms for that particular content embedded content. uh also in terms of motor or mobility impairments. So reliance on mouse inputs. So uh literally traditional technologies are reliant on mouse-based inputs and output interfaces require a mouse for interaction. For example, a screenbased interface which is an output device uh necessitiates the use of mouse or touch screen based interaction where it is again like a cursor which is uh you know moving in the space and is free to touch or free to uh click any part. But then even interaction with with a mouse kind of an interface is is um an issue and uh we have had a discussion about the same in the previous lecture where we spoke about traditional and emerging input technologies. You can go back and check that session as well. So which is a limitation for users who navigate only using a keyboard. Touchscreen sensitivity. So the physical design that the technical specification of some touchscreens may be too sensitive or or very unresponsive for individuals with limited fine motor control or involuntary movements like tremors etc. If it is too sensitive and the person has a tendency to have tremors, there may be multiple clicks which are um uh while uh or on the other hand if the screen is too unresponsive, people are you know trying to press the button but it is not pressing. So the physical design and the physical sensitivity uh specifications of the device is also something that needs to be carefully looked at when looking at motor related impairments accessibility. Then we come to cognitive and learning impairments. So complex interfaces like interfaces that are visually cluttered, lack consistent navigation or use complex languages or forms and that can be challenging for individuals with learning or cognitive disabilities. timesensitive information for example something a password needs to be entered in a certain given amount of time or uh OTP needs to be fetched from somewhere else and then entered in a certain given time so that all of those uh activities I'm I'm sure you are now able to enjoy auto detect of OTPs and uh autofill of OTPS all of those access aspects But yes uh earlier there were it was not the case and then all of these kinds of innovations in the digital accessibility space are not only making the interfaces more accessible for people with cognitive limitations but also for everyone. So again we come back to this same image the same image we saw in our input uh emerging input technologies talk. Um so if because this is the kind of interaction we look at on an everyday basis uh let us now kind of analyze this interface from a output um technology perspective. From an output technology perspective it has a screen. It it has some speakers audio. It may have vibration virotactile right and uh because it's handheld can use that as well. It can emit light uh in cases when the phone is silent etc. So the screen interface there is a lot of um guidelines around it but you know these are basics which need to be thought about uh even for the clickability the navigation the perceivability all of those aspects then audio has to be semantically intertwined with the video screen reader compliance talkback vibrations which are not random which are also semantically associated. So semantics play a very important role uh in uh multimodal input and output interaction. So since there can be multiple output venues uh audio as well as video as well as a vibration it should all be in sync with each other and should have some meaningful semantic association. So let us talk about a simple example. So I'm sure all of you might have used a wallet or for a UPI payment uh on your phone. So right after uh the completion of the payment most of the apps uh you know have that green tick with that uh kind of success kind of um visual and success kind of a sound and some vibratory feedback. So all of these things happen at the same time. So uh it's a multi-ensory input. So if I am hard of hearing I can see the screen that yes the payment has gone through. If uh it is um you know uh if somebody is visually impaired they can hear the audio feedback or the tactile feedback and if somebody's deaf and blind they can you know feel the audio feedback. uh or if you're kind of distracted even then uh you can just kind of rely on the uh virotactile feedback and move ahead. So um let us now move to emerging output technologies. So emerging output technologies are driven again largely by artificial intelligence and are transforming accessibility by offering more personalized uh intuitive and integrated ways for people with disabilities to interact with digital and physical environments. So this is a basic difference between a single model model AI model uh single model AI model. So single type of data, single model and in case of a multi-AI model uh it can take text, image, charts, audio, multiple kinds of inputs and generate multiple kinds of outputs. So multimodal AI is able to enable multimodel interactions. Uh so input modalities can be bio signals for you know say BCI brain computer interactions. Uh sensor based inputs uh so camera all of those aspects count as sensors. camera, mic, pressure sensor, temperature sensor, all of those. Visual signal again which is a camera. Then there's a sound which is a mic. Tangible button pressing which is like an intentional um kind of an interaction and uh the system is able to generate multimodal output. So graphic, sound, haptic, sensor based output as well. So let us look at some examples in the emerging tech domain. Uh so we have classified them based on the um uh kind of domain it falls into. So first is artificial uh a intelligence and machine learning based technologies. So AI is the core engine behind many advancements enabling more sophisticated and flexible assistive technology. So for example enhanced screen readers and TTS uh systems or texttospech systems. Uh so AI powered screen readers and TTS tools. So TTS is read aloud uh tool which I think again we spoke about in the pre in the AI session. So uh even if you now open your system and you see um your your Microsoft Word uh the latest version now has a read aloud u button. If you go in the review it you can press that button and it will start reading the document. And the similarly in the main dashboard there is also a mic logo which says dictate. Dictate is speech to text. So it is an input emerging input technology where it is speech can be converted directly into text. So TTS is now able to provide more natural sound. Slowly it is becoming less robotic and better contextual understanding of the content making the listening experiences less robotic. Yeah. and more seamless for blind and low vision users and those with dyslexia. realtime captioning and transcription. So automated uh speech recognition combined with AI algorithms are able to provide realtime captioning for live conversations and they are also able to do a realtime translation right um so one is listening uh so automated speech recognition listening and understanding and converting it into text of say one language and then translating it in real time into text of a different language and then again communicating it in an audio format using text to speech. So this kind of a system is some something which can you know totally uh be a gamecher in terms of um you know somebody's traveling from um Japan to India and they don't know the language they are able to speak into the system uh in their language and the output is able to in real time understand translate and communicate it uh to a local person in the regional language which is say Hindi. So it can be a very good uh so it can be a very good tool for uh removing such you know limitations like language limitations then there are there can be um you know models which are able to do image and scene recognition. uh so something like Microsoft seeing AI uh it is now able to do realtime reading so image analysis obstacle detection navigation etc reading signs, describing a room, uh you know, attending meetings, telling the uh person with visual impairment like who all are sitting on the table, which person is sitting on what side of you and all of those aspects can be really helpful in the social inclusion um part of so the last point is content simplification. So, generative AI models can simplify complex text into more readable formats and automatically generate descriptive all text for images aiding users with cognitive or learning disabilities. So descriptive all text for images as well as videos can be generated autogenerated and the systems will be able to do that uh by themselves and also especially for longer long format documents uh AI models can simplify the complex text and kind of summarize it for your understanding and make it into more readable formats. So yeah, I mean this is just one example um of live um of a potential AIdriven output technology which is live captioning. So uh you know the there can be somebody who's delivering a speech there can be somebody who's delivering a speech and uh you know the video which is being played uh in the same stadium or the same auditorium is able to have live captioning generated in real time And thus the gathering becomes accessible for persons with hearing impairment because not all people may lo know sign language. And maybe you would feel that why not sign language but not all people would be knowing sign language because it's a different all together a new language right and also sign language is something which is again uh taught and learned by hard of hearing or uh people who are deaf um deaf and mute but for example people who are hard of hearing ing and are not mute. They don't really know most of the time sign language. And also um elderly who are who have uh their uh auditory senses declining, they may also benefit from this technology. Now let's move on to the next category of emerging technologies is which is augmented reality and virtual reality based uh technologies. So AR and VR are moving beyond just you know arcade games and uh now are able to provide practical realtime assistance by overlaying digital information onto the physical world. So navigation guidance is one important use case which can benefit from AR applications. Uh, AR applications can overlay virtual navigation paths onto a user's live camera view, providing realtime audio and haptic feedback to help visually impaired individuals navigate complex indoor and outdoor spaces. Although I would uh say that uh uh navigation of indoor spaces is still a very big open problem and even outdoor spaces like parks eta because there are several aspects to navigation and uh of course AR can be one of the aspects can um cover one of or two of the aspects several uh things like um uh live camera can help in uh automated uh obstacle detection can help in um realtime audio and virotactile feedback in wayfinding uh all of those aspects contextual cues. So AR glasses can provide realtime information about the surroundings such as identifying objects or people or translating sign language into text in real time for an improved communication. Such technologies can offer accessible virtual environments. So for example, VR offers the potential for creating customizable virtual spaces, be it offices, be it museums, etc. Allowing people with physical challenges to participate in act activities that might be difficult in the physical world. And this is like uh one example where um a virtual reality uh interface can allow people to interact with a museum kind of a setting where they're able to walk around and u uh you know interact with u different aspects of um so it can be used for educational purposes. It can be used in tourism etc. Now the third category of emerging input technologies is under advanced human computer interfaces and variables. So innovations in how users interact with devices uh that are creating new output methods beyond traditional scenes is happening at a very fast pace. So um BCI or brain computer interaction is something that we have uh talked about um still in early development but BCI aims to allow individuals with severe motor impairments to control devices uh robotic prosthesis or even communicate using neural signals. Haptic feedback system. So haptic technology in smart watches or variables can use vibrations to convey information. Of course the patterns etc that needs to be studied uh from a research perspective uh such as directional cues for navigation or notification for users with hearing impairments. Uh you know just not relying only on audio cues or visual alerts. voice activated technology. So voice assistants like Siri, Alexa, Google have become mainstream input output tools allowing hands-free operation of smart houses and devices uh in an IoT setting which is particularly beneficial for those with motor impairments or visual impairments. uh you know even if you are able to walk identifying the buttons and pointing the remote at the TV and pressing the right buttons all of that also becomes very tricky but say with Alexa or Google assistants uh just by uh just from where you are sitting you are able to control the television or you able to control your playlist and I think that is something which is of great help. Then uh we come to tactile braille displays. Uh advancements in refreshable braille displays um is um you know as a haptic uh textual display as well as graphical display is uh making a lot of progress. Uh it's now making digital content including graphics and complex equations accessible via tactile output. Um so this is an example image of a tackle graphic display. Uh so maybe now uh you know studying something like biology where you need to kind of go through more and more diagrams every day uh that can be more accessible for persons or students. So I think we uh we saw the same image in the previous session also and the point of it is that we are largely moving from a screenbased desktopbased interface to a very interactive uh kind of an environment uh which will be enabled by a lot of these technologies where it would feel as if you're talking to u um robotic humanoid and uh how you know the other person is able to perceive and semantically process and reply to all your queries and do the action items requested by you. So to summarize the future of accessibility lies in the seamless integration of these technologies into everyday consumer products. Uh it is a still it is still a long time to go but the core principles of accessibility by design uh should be followed rather than thinking of accessibility as an afterthought and u uh some of the key words which you can use for knowing more about this topic is emerging output technologies AI based output devices variable output devices adaptive output devices And uh I think um this kind of technologies uh will have a lot of potential in transforming digital accessibility and digital interaction. uh all the while we have still a long way to go to talk about compliance and regulation about all of these dynamic uh input and output devices because most of the content is generated uh in real time uh and WAG currently has not been designed with dynamic content in mind. It is largely designed for an static interface. So uh guidelines and policies around the same also need to be thought of. And uh in order to know more about the about these gaps, you can refer to uh AI and accessibility lecture part two and uh where we have talked about all of these uh issues in detail. Uh thank you for joining us today and uh we'll see you in the next session.