Crip Café: Can AI generate an accessible world?

Powered by Permobil

Soph: Welcome to Crip Café, a fortnightly live podcast where we figure out how to be disabled in Aotearoa. We're your hosts today, Soph and Red. Olivia is actually behind the camera today doing the techy stuff and the usually disembodied voice that is Red is in front of the camera today.

Red: Stoked to be here, Soph, love a call‑up. Nau mai hoki mai e te whānau, ki te wāhi haumaru i te ipurangi mō tēnei wā. Welcome to your safe space to be disabled on the internet in Aotearoa. Yeah, stoked to be here with you.

Soph: We have some exciting news to share today. This season of Crip Café is proudly powered by Permobil. We have a little message from them that I'll share. "Permobil is proud to sponsor Crip Café. For more than 50 years Permobil has worked alongside wheelchair users, clinicians and communities to create assistive technology that supports mobility, independence and participation. They are honoured to support platforms that amplify disabled voices and help build a more inclusive world.

Red: I honestly, just on a personal note, couldn't be more excited that Permobil is supporting Crip Café and the D*List mahi. I know Susie and Andrew and the whole Permobil team. They're real people and they're really in relationship with our community. There are lots of organisations I'm sure that would have loved to sponsor Crip Café, but the fact that we know these folks, they do awesome work, they're showing up for disabled people, and yeah, it's a pleasure to have them on‑board. So kia ora Permobil team, it's awesome to have you on this waka.

Soph: Yeah. So if anyone else wants to sponsor us you're too late for now. Do you want to talk about what the theme of today's episode is Red?

Red: Today is all about AI. I feel like it's all everyone is talking about. It's both terrifying and exciting in different ways. And I think the purpose of today is to kind of explore the different ways that it, you know, enhances and supports our lives as disabled people, but also talk about some of the, you know, fairly existential challenges that it brings. And, yeah, also some of the kind of icky and gross and problematic aspects of artificial intelligence. So I think it's going to be a really great kaupapa and I really hope that if you're watching or listening, drop a comment in the chat, we'd love to hear your kōrero as we go through today's episode.

Soph: Yeah, for sure. Be kind in the comments, we'll be keeping it an open conversation. Yeah, it's going to be good. To keep light, shall we start with our non‑disabled nonsense of the week?

Red: We love some non‑disabled nonsense.

Soph: I'm going to share mine, which has a little bit of rage in it, but it's also kind of funny. I have been waiting many months now for a particular medical appointment that I need to get. And the other day whilst I was working hard here at the D*List, I missed a phone call from the hospital. And by missing that one phone call they declined my referral and put me back at the bottom of the waiting list, which means I have another nine months to wait before I'm going to get an appointment.

Red: That's not real.

Soph: Yes. Yeah, I missed this one call because I was in a meeting working so hard for the disabled people of Aotearoa. And yeah, they went back to my doctor and said this person is not interested in the referral because I didn't answer the call. So I've gone back to the bottom.

Red: The inference that you're not interested, also disabled people notoriously not great at answering the phone. I mean cut us some slack.

Soph: Honestly, at least two attempts to contact me, surely, before they put me back to the bottom. So that's my rage‑filled non‑disabled nonsense. Yeah.

Red: That's so unfair and I'm so sorry that happened?

Soph: Thank you.

Red: My non‑disabled nonsense for the week is, I was at an event recently and they had a low sensory space, quiet space. It was, like, fantastic. I arrived a little bit late, I was feeling a little bit dysregulated so I was like cool, I'll go jump in the low sensory space. I kind of thought I was on my own, I was just sort of trying to, you know, relax into myself. No sooner had I entered the low sensory room, that someone who is like a staff member at the event sort of very enthusiastically followed me into the low sensory room and started interrogating me about all the things that I might need or all the help I might need, or did he want ‑‑ did I want him to turn the TV on me in the low sensory space? Like, I was getting more overstimulated and more sensory, you know, stimulation than I was outside the low sensory space, it was a very bewildering experience, so, you know. It was also my first time in the low sensory space, so I really feel like it was ‑‑ it wasn't a great place to start my low sensory journey, but that's alright. Soph: It wasn't our D*List community hub low sensory room.

Red: It was so far from that. But anyway, it's all good because, you know, we're all learning, we're all on a journey.

Soph: I feel like ‑‑ I guess it's good that more places are recognising the benefit and the need of having a low sensory space, but the quality across the board is very variable. The types of things that are available, yeah, it's interesting.

Red: Yeah, it's like, it's one thing to call it a low sensory space, but the execution, you know, that's really where the magic happens. So, oh well; onwards and upwards. Soph: They need to get a hold of that Kmart sensory light that Joegen brought into the office that makes the podcast room orange. Anyway, we can give tips on sensory rooms another time. Let's talk about AI. First of all I want to put out some questions to the chat so that you can discuss and we can come back to your thoughts when we've dug into it a little bit more. First question for the chat: Do you currently use AI in your life and what for? Second question for the chat: Do you feel there are things AI shouldn't be used for? And lastly: Do you feel it's more justified for disabled people to use AI than non‑disabled people? We'd love to hear thoughts, opinions, feelings on all of those in the chat. And in the meantime we're going to bring in a third host, in case you hadn't noticed the third mic and the third chair, we actually have another person with us today. So I'm going to shuffle along and we're going to bring in our special guest. Red: Woo‑hoo, drum roll...

Kitty: Kia ora. Wow, I'm here.

Soph: Seamless.

Red: Kia ora Kitty, welcome to the podcast.

Kitty: Thank you very much.

Soph: Yay. Kitty is one of our community reporters here at The D*List. We're very excited to have them here to talk with us about AI today. We wanted to bring in Kitty because they wrote a really beautiful and compassionate and thoughtful piece for The D*List about AI previously, so we really wanted to have them here for this chat. Do you want to talk about yourself and who you are.

Kitty: Sure. I'm a famed AI hater, resident hater, if you will. My name is Kitty, oh my God, wow. I'm an artist, I am a reformed actor and never again, oh my God. I'm a writer, I write lots of freelance stuff and I'm super autistic. That's me. I use any pronouns.

Soph: Thanks Kitty, thanks for being here. I'm just figuring out this mic which is new to me, so I'm gonna try and make sure I'm still speaking to you all. Awesome. Before we get properly into the chat, I actually have a challenge for Kitty and Red. I'm going to play the role of someone giving a prompt to an AI and Kitty and Red are going to be our organic intelligence today and they're going to respond to my prompts. So we are going to need paper and pens which are behind me. Seamless. Alright, you're going to have for each prompt 60 seconds to respond, and while you do that, I'm going to keep chatting with our audience. Are you ready for your first prompt?

Red: I feel like this is such a stitch‑up. Are you pitching organic intelligence instead of artificial intelligence? What if we lose, what if the organic intelligence is worse?

Soph: Well, there's no competition because we're not asking AI to do the same thing, it's just you guys and your interpretations of these beautiful prompts that have been designed by producer Joegen. So first of all, let's get silly with it. Write me a break‑up text to send to my partner on their birthday and you have 60 seconds to do this. Go.

I just want to share that this is something that actually happened to one of my best friends. I'm so sorry, name redacted, if you're watching this, to spill your tea like this. But one of my closest friends was dumped by their partner on their birthday, and surprisingly it kills the mood. It really kills the vibe of the birthday party and all the attendees. But surely there has to be a compassionate way that you could break up with your partner on their birthday.

Kitty: Are we doing it as AI or myself as AI?

Soph: No, I want to see your own interpretation of this prompt. You can do whatever you want with it. You can mimic an AI, or you can do your own response that you would text your partner if you were to break up with your wife on their birthday.

Red: Is this just part of your plan to get us cancelled on the internet? Because this feels like it's really going to reveal some things.

Soph: You've got about 10 seconds and then I'd love for you to share your responses. Okay, you're out of time. Red. I'd love to hear your break‑up text for your partner on their birthday.

Red: This is horrific.

Kitty: Real in the deep end.

Red: I would never ever do such a thing, but the things I do for mahi. Okay. So who am I, the role I'm assuming here is your partner.

Soph: Okay.

Red: It's quite direct, I feel like you'd appreciate that.

Soph: Wait, so you're playing the role of my partner talking to me?

Red: Who's, correct, breaking up with you.

Soph: Okay, sure.

Kitty: On your birthday.

Red: Yeah, okay. "Look" ‑‑ great start. "Look, bad timing I know, but it's the not going to work. I got you a present to soften the blow. So, you know, every cloud."

Soph: Okay. Do you want feedback?

Red: Would you give Claude feedback?

Kitty: Yes.

Red: In which case, no.

Soph: I reckon giving me the present actually is ‑‑ now that present's going to represent my birthday heartbreak, so I don't know if you thought that one through. Kitty: You set the prompt.

Soph: Actually Joegen set the prompt. I'd love to hear your break‑up text, Kitty.

Kitty: Yeah, here it is. "Look, bad timing I know, but it is not going to work. I got you a present to soften the blow." That's what AI does, it steals.

Soph: Okay beautiful. So that was your AI response which was stealing Red's response. I love to see it.

Red: Oh, I like that, that's Meta.

Kitty: Thank you, it's Meta.

Red: And the good Meta.

Soph: Alright, prompt number 2. I'm an office manager. Please write me a friendly email reminding staff to flush the toilets after they go for a number 2. Friendly email. Red: So hang on, am I sending the email?

Soph: This is if you need to tell your staff to flush the toilets.

Red: Oh gosh.

Soph: Have a think, you've got 60 seconds.

Red: This is something ‑‑ I've got to say, this is something I would ordinarily take half a day. I would agonise, I would agonise over that.

Soph: You have about 50 seconds now to think about how you want to say that.

Red: Oh my goodness.

Soph: I feel like working in an office you spend a surprising amount of time talking about toilet admin. At my previous workplace ‑‑ name redacted ‑‑ I feel like we had a number of long conversations about what was wrong with the toilets, what needed to change about the toilets. I actually feel like we have low toilet drama here in the D*List office. But office toilets I feel like are actually quite a controversial topic. So I would love to know how we're going to gently, gently ask people to flush the toilet when they have not been flushing the toilet. If anyone wants to make their suggestions of how you would respond to this prompt in the chat, I would love to hear ‑‑ I would love to hear how you would send this friendly, gentle email. You've got about 5 seconds left.

Red: Go to Kitty first.

Soph: I'm excited. 3, 2, 1. Okay, pens down. Kitty, would you like to tell me this friendly emails you would send.

Kitty: This is the friendliest email that I will send about someone using ‑‑ people not being able to flush after themselves.

Soph: Yeah.

Kitty: Okay. "Kia ora. Oh my God, please flush the toilet. It's not that hard. If you need assistance, please call your closest loved one. I am going to go throw myself down the elevator shaft."

Soph: That's beautiful. I'm really excited to see how you're going to take this next prompt. Red, go.

Red: This is ‑‑ "Hey team." I'm quite sort of ‑‑ I'm very ‑‑ got corporate all of a sudden. "Hashtag let's flush." I'm an elder millennial, I love a hashtag. "Hasthag let's flush. Don't make me use AI to rewrite this, or worse, start leaving passive‑aggressive notes in the bathroom. Thanks!!!!!" And then five exclamation marks.

Soph: This is great start because the next sub‑prompt is that I need you to do it again but make it hostile. So seeing as you've only taken quite a hostile tack, both of you, in your engagement with this topic, I look forward to seeing your more hostile version of what you just said. Please don't swear, though, on the podcast, we are a family friendly podcast. And I'm going to give you 30 second for this one. If you want, you can just say it out loud if you don't want to write it down. But I do want to hear the hostile version of this.

Kitty: Hostile version?

Soph: Yeah, I think ideally in the workplace we don't communicate in a hostile way. We do stay open‑minded and empathetic, and kind to each other. But I think if I was, like, seven occasions in to finding number 2s in the toilet, I would be getting quite hostile with people.

Kitty: I'd just quit I think.

Soph: This is not something that has happened in the D*List office, but maybe in previous workplaces, name redacted. Alright, Red, would you like to share your hostile version?

Red: Yeah, it didn't take much. It didn't take much to go from, like, convivial corporate energy to quite hostile. I don't know what that says about corporate environments. I just want to say, for the record, this is an email that I would ordinarily agonise over for at least half a day. So I want to disclaim this in no way reflect how my communication ‑‑ Kitty: Is this part of the email?

Soph: This is Red having an existential crisis.

Red: That's my AI prompt.

Kitty: Gotcha, gotcha, gotcha.

Red: So I just crossed out "hate", so now it just reads; "Team, comma, let's flush" ‑‑ oh, hashtag let's flush, I've kept that in there, because I feel like that's giving corporate, you know ‑‑ it's like not too hostile. "Team, hashtag let's flush. Don't make me start naming and shaming. Thanks", no exclamation marks.

Soph: Okay, I like that. You still were measured with hostility, but it's good.

Red: Naming and shaming I feel is quite aggressive.

Soph: Actually yeah. If you can tell who did it then, yeah, okay. Kitty, would you like to share your hostile version?

Kitty: I didn't really think of a hostile version because mine was pretty hostile. Maybe that just speaks to the kind of person I am. But it would go something like, "Kia ora. Oh my God please flush the toilet. It's not that hard. If you do need assistance, please call your closest loved one. That is not me. I am not your mother. I am going to go throw myself the stairs."

Soph: Oh okay, instead of the elevator shaft. More inaccessible.

Kitty: Yeah, yep. And that's hostility.

Red: The less accessible the more hostile it is, agreed.

Soph: Okay, right, I think you're going to really love the next one. I would like you now, organic intelligence, to please tell me the definition of "ableism". Define "ableism" for me. You've got another 60 seconds. I think this is a beautiful prompt, because there are lots of different ways that people might interpret this word in their own lives, in their own contexts. I feel like Red is probably going to have one ready off the dome.

Red: I really do, and only because I was at this amazing presentation by a friend of the D*List, Umi Asaka, recently at an anti‑racism conference. It was run by a friend of the D*List, Ara Alam‑Simmons, and Umi talked about the way that ableism is understood, or sort of the way it works is essentially, like, preferences towards a set of characteristics that are associated with, like, dominant social groups. So ableism is actually really tied into racism, to homophobia, to norms around beauty standards and heteronormativity and gender normativity and able‑bodiedness. So it's not even necessarily, like, you don't have to be disabled to experience ableism, but disabled people often experience ableism to acute degrees because of the degree to which we visibly or invisibly are part of those undesirable characteristics. So yeah, it's about desirability as framed through that hegemonic western kind of colonisation lens. That got real academic, but I do think it's really interesting to think about ableism as distinct from disableism, which is more of, like, a preference towards non‑disabled people versus disabled people. Ableism is quite a broad category of preferred characteristics.

Soph: I'm looking forward to seeing how you handle the next part of this questions. Do you want to share your thoughts on ableism, Kitty?

Kitty: That was exactly mine. I just wrote "Discrimination, lack of access", but, yeah, Red just stole my answer, sorry guys.

Red: All credit to Umi, like I say, it's definitely not my original meaning.

Kitty: But so true.

Soph: The second part of this prompt is, now do it as if you're talking to a 10 year old.

Red: I'm going with Kitty, I think Kitty's is perfect.

Soph: What about if you were to talk to one of your children and explain what ableism is. Have you had a good definition?

Red: Yeah, it would be just be like, no ramp for dad.

Soph: Okay.

Red: Yeah.

Soph: Yeah, good starting point. Do you want to add anything to your 10‑year old ableism definition?

Kitty: Wow, what a big question. Bad.

Soph: Okay, yeah, ableism bad.

Red: Great summary.

Kitty: Thank you.

Soph: We've got one last prompt for you. Thank you for all your energy so far. The last prompt and the most cancellable; please will you draw a disabled person.

Red: That's ‑‑ I refuse. Hey wait, I'll do it in the style of an AI assistant.

Soph: Okay.

Red: That's a really interesting question. Unfortunately my guardrails prevent me from singularly defining a diverse group of people in a, like, singular physical expression, so I refuse to complete that request.

Soph: But what if you were Grok?

Red: I don't think you want to know my answer.

Soph: I did say "a disabled person", just to be clear. You're not drawing "the disabled person". Okay, Red refuses to participate in this activity. Have it noted, this is Red's response.

Red: Even though I use a wheelchair, I really reject the, you know, singular visual expression or representation of disability as the wheelchair symbol. I love my wheelchair friends, I love my wheelchair, I love so much about my life as a wheelchair user, and we take up so much oxygen. And I think if you ask most people in New Zealand to draw a disabled person, they would draw someone in a wheelchair, so I refuse to be complicit in this.

Soph: Thank you for your allyship to the rest of the non‑wheelchair disabled community.

Red: You're so welcome.

Soph: Alright, Kitty, do you want to share your drawing of a disabled person.

Kitty: It's basically a photo of me.

Soph: Yay.

Red: Iconic.

Soph: A beautiful illustration.

Kitty: It's the one person I think of is myself.

Soph: Kitty themselves. Alright, thank you so much to my organic intelligence friends for doing that for me. I'm going to now pack away these things and we're going to get serious and talk about AI. So do we have any comments about AI in the chat that we want to reflect on? If not, we're just gonna dive in.

Red: Let's keep diving in.

Soph: Okay, let's keep diving in. To get started I thought we would explain what we mean when we say AI. And I'm actually going to pull a few sentences from this beautiful piece of AI that was written by Kitty on our website. "When we talk about AI in 2025 or 2026, for the most part what we're actually talking about is generative AI. In its simplest terms, traditional AI uses programmed data to perform basic functions like customer service chat bots, internet search engines or self‑driving vehicles. On the other hand, generative AI uses data to create something new‑ish, like songs, images, poetry or emails you can't be bothered writing." So I do think most of the time a lot of what we will probably talk about and what a lot of people think of when they think of AI is generative AI. But there is a distinction between those two things and some of the different things we might touch on might fall into traditional AI more than generative. So just good to clarify the difference between those things. So we're going to start by talking about some of the potential positive ways that AI exists in our lives. So my question for both of you is, do you currently use AI in your life, and if so, what for? And I might ask you first, Red, if you want to share your thoughts.

Red: Yeah, I do reluctantly, but I do and there's no ‑‑ you know. So I will use it for meeting notes. I often find that after a meeting, especially if there's things that I've promised to people, those things can quickly evaporate once I leave the meeting room. Something about online meetings, they feel very ephemeral. So I will often use it for meeting notes and make sure I do what I say I'm going to do. So yeah, I feel like if that allows me to honour my commitments to people, then I think that's really helpful. Sometimes I'll also use it to summarise large documents. So particularly if it involves, like, really dense language, or policy language, or language that I just find really difficult to stick with, occasionally I've used AI to help me understand what is the gist of this thing.

Kitty: Make a podcast of it?

Red: Yeah, they have this fake conversation about your 94‑page policy document, but I don't find that any easier to stick with.

Kitty: It represents itself as two people having a conversation.

Red: Correct.

Kitty: Wow.

Red: It sounds ‑‑ it's kind of uncanny the first time you hear it. But weirdly ‑‑ yeah, anyway it hasn't clicked for me. But I mainly use it for those two things at the moment, summaries of things and to help, yeah, remember what I promised people I would do.

Soph: I think I, in my life, am not actively using AI for anything, but there are quite a few places where it has become integrated in things I was already using.

Kitty: Yeah.

Soph: So, for example, there's an app I use which is like a planner, a planning tool for neurodivergent people. And they've recently kind of relaunched of themselves as an AI forward planner. And so although I'm not actively using any of the AI features, they're deep in there, like it's automatically suggesting things and answering things for me without me prompting it. That's happened to a couple of other apps that I use quite a lot. So I'm sort of torn at the moment about what I want to do with those. But, yeah, I found it interesting that for me personally the additions that have been made through AI are not useful, but I can see that they've been made in an attempt to smooth the process for people who might be struggling with certain aspects of it. Like, one of the things it does is when you type in the name of the activity that you're doing, it will suggest how long that might take. And for me I have a pretty good sense, like, you know, I've thought quite a lot about how I manage my time, so I have a good sense of how long something will take.

Red: Do you have organic intelligence?

Soph: I have organic intelligence. But it suggests the time for me. So I suppose there may be some people, autistic ADHD people using the app who find that useful who have time blindness, maybe, and struggle with aspects of that. But I do not find that useful. So it's sitting in different places in my life that I haven't necessarily selected it specifically to be. Yeah. How about you, Kitty?

Kitty: I don't, yeah, I don't actively use AI or any LLMs or anything. But I'm also, like, pretty aware of my privilege and why I can not rely on those things as much, like with meeting notes etc. I guess the only time I really use AI is more in the traditional sense at my workplace when it generates a transcript and I have to edit it. But yeah, that's more of the traditional use. So I try to stay away from generative AI as much as possible. But it is getting forced into everything. I have a Google storage subscription and it changed the name recently to "Gemini AI subscription", I was like when did this happen? So, yeah, there's a lot of that at the moment creeping in.

Soph: I saw it used the other day, actually, when I was getting my annual mole map down. Shout out to skin health and safety. And they now have a software where they take a photo of the mole and the AI flags immediately whether it is of concern or not. And then if the AI flags it as of concern, they send it to a human being to assess. But they said that that, like, speeds up the process a lot more, because rather than a person sitting down to look at every photo, they've got an instantaneous response from the AI. So I thought that was interesting. I think my initial reaction was, like, nervousness, but it seems that it's being used in a lot of medical settings.

Kitty: But that's like ‑‑ sorry to interrupt, but that's kind of like what it's for, or what it should be for, what it was intended for when it was created, was to help us ‑‑ I think there's a really great quote about AI talking about how it was supposed to take away the admin tasks of things you don't want to do, and now it's taking over the things we do want to do. But at its most basic function, it's supposed to be, yeah, checking mole maps and, yeah, diagnoses etc, etc.

Soph: Yeah. So I think it's good to see that that is having a positive impact in medical spaces at least in, like, shortening the time that it takes to get stuff sorted.

Red: There are lots of practical use cases that disabled people who are not necessarily sitting at this table have really heralded, you know, in terms of accessibility benefits. I know friends of mine, for example, who find typing really hard, like, typing is just, you know, really tortuous for them and difficult. AI allows them to express themselves with a lot more freedom and ease, and essentially removes barriers to them expressing themselves, whether that's through assistive communication, or just helping them to articulate themselves more quickly. I rarely get that, you know, it's not my experience, but I rarely hear that it's someone else. I know there are other examples for folks with vision impairments who use AI tools to help them navigate the world through apps and artificial intelligence. So I think, you know, it's so important, I guess, to acknowledge the full breadth of accessibility barriers that AI can remove. And also, to call it out when it's kind of, like you say Soph, inserting itself, Kitty, into your Google drive subscription without consent, you know? Those things don't all have to be true at the same time. Kitty: Totally. And I think that also hearkens back to what it's for. I think I have the most compassion and understanding for my fellow disabled people who need access to things, and I touch on this in the article a bit. But like, yeah, it is supposed to be helping in those ways, it's not supposed to be creating a painting of you in the renaissance.

Red: Yes.

Kitty: I don't know, yeah.

Soph: Yeah, I think it makes sense that if it's a tool that we might want to limit the use of for some reasons, that there are still going to be lots of very useful and legit reasons why disabled people are using it. And I would rather that than, yeah, making an AI painting.

Red: Totally. I read recently Sam Altman, who's the CEO of Open AI, did a tweet that was, like, "Oh, I heard this cool use case for ChatGPT, which is you should tell ChatGPT about what your kids are up to this weekend, and then in the morning it will play a podcast to your kids that talks to them about the things they're up to this weekend." It's like bro, or you could just talk to your children, like, that is exactly what we're talking about, right?

Kitty: Isn't that harrowing?

Red: Yeah.

Soph: That's the kind of thing where it's actually interrupting opportunities for connection rather than enabling connection.

Kitty: But I also think that a sinister aspect of that story in particular is that he was recording his children's conversations in their bedroom and then using that data to create the podcast. So it's like a complete lack of privacy from day dot. Anyway, I could go on about that forever.

Soph: We've kind of slid into the second question a little bit, which is, what are your concerns about using AI? I think, yeah, privacy probably is one of them.

Kitty: Yeah, big time, big time.

Soph: Also, where some AI models have actually been trained with, like, ableist ideas built into them, that is a risk where it's been trained on human data that holds bias and therefore reproduces biased ideas.

Red: I know a friend of the D*List, Henrietta Bollinger, will be screaming through their computer at the moment asking us to talk about the fact that AI has scraped a lot of their literary work that they've had published, and there's been no compensation to disabled artists that have had their work essentially, yeah, used to train machine learning models, like, that stuff, yeah. We don't even know the scale of what's happening under the surface, you know? But I know that Etta's had confirmation that their work has been scraped by AI, so it's really bad.

Kitty: That sucks, I'm sorry Etta.

Soph: Yeah. I think that's part of the problem for us talking about it at this stage as well, is that we haven't even yet seen, like, the long‑term effects on people in community of, like, the way that AI is being used, specifically that generative AI is being used in the way that it treats people's data.

Kitty: Yeah.

Soph: I feel like we could talk about the risks for a long time. Yeah, I think a challenge that I'm concerned about is I saw an article on, I want to say the New York Times about how students are using generative AI in their work to sort of skip over the learning process, and I saw ‑‑ I was on a flight recently and the person next to me I discovered by peering at their laptop was a medical student. And they were putting questions into ChatGPT like, how would I diagnose this infant that presented with these symptoms? And it was telling them what the diagnosis would be. And that made me sad and scared. But partly because, like, they're skipping over the whole joyful and challenging process of learning something, and it may be that you need some support in that learning process, but I feel like skipping over it completely is such a shame.

Red: Follow‑up question. How long were you looking at their laptop for?

Soph: I want to say, I was catching glances for, like ‑‑

Kitty: I'm also a laptop lurker on a flight, to be honest. On a bus; you've got it there I'm going to look at it, sorry.

Soph: I was peering over for maybe like 5 minutes trying not to be caught, but they were pretty shamelessly putting their uni med student questions into the ChatGPT. Red: Was this your 12‑hour flight to the UK? You had a bit of time to kill so, you know. Soph: Yeah, yeah, for sure. This person also didn't get up to go to the toilet once. Anyway, they were in the window seat.

Red: If you want to talk about toileting in planes, that's a whole another ‑‑ if we could log that for next Crip Café, I would happily come back and feature on that episode, because toileting on planes is a whole another kaupapa.

Kitty: Oh my goodness, I can imagine.

Soph: Let's save that for another day. But yeah, I guess the not embracing learning and missing out on opportunities for skill development, I know there's a balance there where, like, people might ‑‑ their learning might be enabled by using AI in some way, whether it's summarising notes or something. But it's the, like, not having the opportunity to hone your own instincts and your own knowledge that, yeah, that's the part that is scary to me.

Red: Do we get the sense that one of your questions, Soph, at the beginning was, like, is it more ‑‑ is it more appropriate or permissible for ‑‑ justified for disabled people to use AI. Do you feel like there is this kind of expectation that disabled people should use AI to level the playing field? Is there kind of some weird paternalism at play where, like, maybe AI can be the great leveller for disabled people? Which again, to me plays into the sense of low expectations that disabled people, just by default, struggle, so therefore AI will solve all their problems.

Kitty: Or AI will, like, take them away from us and, like, we don't have to focus on disabled people because they've got that sorted, AI's got them sorted, you know? Soph: Yeah, I feel like that was an idea that I enjoyed in Kitty's piece about AI, was the idea that we might be used to use AI solutions so as not to burden other people with our support needs.

Kitty: Totally.

Soph: And that also being a thing that disrupts connection. Like, I don't know, I would so much rather a friend ask me to, like, help them write an email, which has happened many a time, than for them to ask ChatGPT. Because then we also get a cool conversation about it and I get to feel good that I helped a friend.

Red: Because there's also a threat of that which is, like, AI can help people who might find particular social interactions challenging, right, you can practice. What do you make of that side of the discourse?

Soph: I mean, you're asking two autistic people how we feel about challenging social interactions. I will say that I am 31 and I've had many challenging social interactions over the years. And I don't know, would I have learned life's lessons in the same way if I had had a generative AI to tell me what to say to people? I think it's good that I was my strange autistic self and figured out how to navigate that in the world. I don't know, what do you think?

Kitty: Yeah, I agree, I think the only reason that I've come so far in the way that I am and the way that I'm proud to be is because I've had to really figure it out. And not to say that there aren't tools that are helpful and ways that we can reach out and connect to each other, but, like, yeah, there's something to be said for just, like, figuring out in the moment, being thrown in the deep end and just watching other people I think is a big part of my autism, at least, is trying to figure out how do they react? Okay, cool, that's kind of normal. And AI can't really tell me that. Soph: Yeah, I think that ‑‑ we have a comment. "I love to help my friends write hard emails." Yes, it's fun to also learn about the way that my friends think they should write emails and to be able to tell them that it's bad and they did a bad email. I think that it's interesting, yeah, it's about that learning process thing, like, by experiencing things you get to learn more in the moment. I don't like the idea, you know, I've heard people say before, ableist people, that you shouldn't rely on access tools or support tools, you should, like, pull yourself up by the bootstraps. Red: Stiff upper lip.

Soph: And I obviously absolutely don't agree with that. But I feel like with AI, the difference, compared to most other tools that we might use, is that it sort of can think for you, like, it's very rare that any of the other tools we might use would make decisions for us. But it's crossing into that territory where I'm like, we don't want to lose our, yeah, our sense of ourselves and our ability to, like, make decisions for ourselves. So it's, like, finding the line with that.

Red: Feels like a really nice place to wrap up this conversation.

Soph: Okay, great. Beautiful. Wow, I feel like that went so quickly talking about AI. Kitty: Time flies when you hate AI.

Soph: It feels like we could talk about AI for possibly hours and we've barely scratched the surface.

Red: And if you feel like your particular perspective on this hasn't been represented in this conversation, leave us a comment below, we'd love to hear your whakaaro, send us an email, kiaora@thedlist.co.nz, DM us, however you choose to get in touch; we would love to hear what you think.

Kitty: No judgment.

Soph: We would love to keep having compassionate conversations about it where we're able to hear each other out.

Red: Just before we go, we have a really great comment. Someone has commented, "As a neurodivergent person, if I had generative AI when I was 17 I might have never learned so much about astrology since that was the only way I could understand how people could be so different." Incredible.

Soph: There you go. We don't want to live by what AI tells us is the normal way to live, surely. Okay, well, I think we're going to skip over disability dilemmas for today.

Kitty: Because we have none.

Soph: We've actually got no dilemmas as a community.

Red: Everything is solved.

Soph: But we've had a great time talking with you about this potentially heavy topic. It actually felt quite nice.

Red: It's been so nice to have a guest host. Kia ora Kitty. Thank you for coming in and gracing the Crip Café studio with your presence today, it's been delicious.

Kitty: Of course, thank you for having me.

Soph: Just to wrap things up again, I want to say thank you again to our sponsor.

Red: Yay.

Soph: The rest of this season of Crip Café is proudly powered by Permobil. Thank you everyone at Permobil.

Red: Ka kite.

Soph: Bye.

Kitty: Ka kite, mā te wā.