🦆 Interaction Nerds by UXA
This week from Archive & Collection: lack of diversity of voices in AI, accessibility of CAPTCHAS, learning from child-computer interaction, and visualizing colorblindness.
Table of Contents:
👋 Announcements — Last issue of the semester and good luck with finals!
🎯 Resource — Which lessons from child-computer interaction should we bring to the rest of UX?
👋 Announcements — Last issue of the semester and good luck with finals!
We hope that everyone is getting prepared for the final stretch of the semester! As finals week approaches, don’t forget to take good care of yourself both physically and mentally while working hard to prepare for your exams and final projects! 💪
As the semester wraps up, this will be our last issue of the semester. Thank you for being around and exploring the different sides of UX and HCI with us, and we will see you soon again next semester! 💗
🌾 Opinion — Why is JARVIS the only male AI?
💬 Siri, Alexa, Cortana, Google Assistant… why are all of these voice assistants female (by default)? Common explanations are that we prefer female voices and that there’s more female voice data out there. Regardless, having primarily female voice assistants is problematic.
There are no social norms with a voice assistant. You don’t have to be polite or say please; you can talk however you wish with zero repercussions. And if children grow up interacting with female voice assistants this way, it perpetuates the bias that
“women are subservient, best suited for secretarial roles wherein they never talk back or protest.”
👀 Whenever AI is male-presenting in media, he becomes a complete character with his own story (eg. JARVIS → Vision in Age of Ultron).
In media, how often are female-presenting AI more than just helpful assistants in the background?
Rather than doing whatever’s easiest to build with existing biased voice data, we need to do what’s right: build AI voice assistants with a diversity of voices.
🍀 Industry — It’s about time CAPTCHAS become accessible
We’ve all been there. All you want to do is access a site, but you keep getting the CAPTCHA wrong, even though you’re a human, not a robot (or at least you think you are)!
As CAPTCHAs get increasingly difficult, even people without impairments are having difficulty solving them. What about people with impairments?
👀 There are 284 million visually impaired people & 39 million blind people in the world. For them, reading an image of warped letters can be impossible.
📕 20% of the population has dyslexia, making word CAPTCHAs painful & frustrating.
🌏 And what about non-English speaking web users?
Here are a few (imperfect) ways to make CAPTCHAS accessible:
❓ Text or logic questions: “What is the 4th letter in the word “heartfelt”? These are typically specific to a language (usually English) & inaccessible for those with cognitive disabilities
📫 Email / SMS verification: effective, but very annoying for users.
🍯 Honey Pot: text fields only seen by bots → if data is inserted, it’s most likely a bot Also traps screen readers and subsequently blind individuals.
🔉 Audio & Visual: user can identify a photo or just type what they hear. With audio, text, and image recognition, bots no longer find these difficult.
❓ What makes a human, human? What makes you different from a bot or AI model?
🎯 Resource — Which lessons from child-computer interaction should we bring to the rest of UX?
Child-Computer Interaction (CCI): the design, testing, and creation of interactive computer systems for children.
The nature of researching & testing with children results in numerous practical, methodological, and ethical considerations taking place at every step of the process. Read on for why CCI research & testing is as robust as it is and how we can apply such procedures to our own research.
Key Takeaways:
🆎 To choose a research method, CCI researchers consider every aspect of their research targets before choosing an approach. The same handful of UX research methods are widely talked about, but there are countless research methodologies; UX researchers should also deeply consider their situation before choosing a method.
🚼 Children are great testers! Kids are curious, have highly expressive body language, and can be quite vocal about their thoughts & opinions. Children notice issues that an adult may never see.
🆗 Detailed ethical considerations ensure that child participants give informed consent and are benefited by the research. Research participants of any age deserve the same careful consideration.
🚸 Kids are NOT mini adults: they have separate perceptions, brains, and social norms. In research, do not assume that one user group is just an altered version of another.
🎯 Resource — Visualizing Colorblindness
🎆 Fun - Random Cool Things!
Who’s behind the scenes?
Thanks for reading this week’s Interaction Nerds by CMU UXA! The editors behind this work are Alana Wu and Rebecca Jiang.
🥰 Follow Us
We look forward to co-curating future content with you! Let us know what you think about the content and send us some good vibes through these channels below.
Instagram | Facebook | Linkedin | UXA Website
🤩 Opportunities
Looking for participants for your research? Wanting to recruit students to work in your lab or on your project? Advertise on the UXA newsletter! Fill out our opportunity form with some information about what you want to advertise, and we'll feature your project/opportunity on the UXA newsletter!
🌱 Join the Community
Join the Slack channel to get updates on upcoming events, learn about cool resources and opportunities, ask questions about anything HCI, and much more.