March 19, 2024 Smashing Newsletter: Issue #448
This newsletter issue was sent out to 206,386 subscribers on Tuesday, March 19, 2024.
Editorial
It was April 29, 2005, when Apple released the fifth and latest iteration of macOS. This was before macOS was even macOS. No, no, it was Mac OS X Tiger, and it retailed for the handsome sum of $129.95. (This was even before the realization that evergreen OS releases at no cost were a faster path to consumers churning through devices.) Free updates wouldn’t come until Mac OS X Mavericks, the system’s tenth major release.
Mac OS Tiger was the first of its kind in a few ways. For example, it was the first Mac OS to come packaged as a DVD instead of a CD (which could be requested instead for an added $9.95). As far as tech goes, though, Tiger was the first version to officially offer VoiceOver, a new accessibility feature announcing content to users who had difficulty reading off a screen. (Mac OS 10.3 Panther sported a preview copy under the name “Spoken Interface Preview”.)
Fast forward with me. It’s March 2024, and OpenAI unveils its latest feature for ChatGPT, its AI-based app that responds to user queries in a conversational format designed to replicate human interactions. ChatGPT gained the ability to read its text aloud to the user in a computer generated voice, not totally unlike Siri or Alexa… or VoiceOver. We got a hint of it in September 2023 when two-way conversation made its debut, but the March release combines it with text generation for the “best” of both worlds.
There are truly impressive things about ChatGPT’s Read Aloud feature: multilingual support, rewind/skip, and whatnot. But it also feels a lot like a solution to a made-up problem. We already have plenty of assistive technologies that are capable of reading text out loud. Not to mention, they are freely available and often bundled right in the OS, a là VoiceOver.
This dovetails nicely with Jeffrey Zeldman’s “Where the people are”. In it, Zeldman reflects on a client project that called for a social networking feature. He could choose to build it out or leverage Facebook’s existing capabilities, preferring the former to the latter because it could offer users convenience and a more branded experience. His humility, however, pushed him toward the latter because Facebook is where the people already were.
Is it impressive that ChatGPT can read things aloud? Yes, but is that where the people already are? There are always two sides to a coin. At what point, though, is this a “damn the torpedos” act of development for development’s sake? Where’s the human factor in the process of re-engineering human interactions? Perhaps humility is the missing piece?
On another note, the Smashing team is working hard on new online workshops. In fact, Vitaly will be running one in July on “Designing For Complex UIs” where he’ll dive into ways to design better AI experiences for customers while respecting digital sustainability and inclusive design at the same time. We also have SmashingConfs taking place in Freiburg, New York, and Antwerp later this year!
We also have friendly bundles for great teams, so perhaps get your incredible team together — for a friendly, inclusive, and memorable experience this year. We can’t wait to see you!
— Geoff Graham
1. AI Interaction Patterns
AI is here to stay. So, how are designers responding to it? When Emily Campbell started consulting teams on AI and product strategy, she noticed how inconsistent the existing experiences already were. Companies are prioritizing “doing AI” but struggling to gracefully integrate it into their current products, so her observation. To help designers make AI feel familiar and straightforward for users, Emily created The Shape of AI.
The Shape of AI is a growing collection of AI interaction patterns that rely on proven human interaction and cognition frameworks. They cover everything from identifiers for distinguishing AI content to different ways of entering prompts and trust indicators for controlling the response and assessing its accuracy.
Another interaction pattern library that helps designers tackle common AI challenges comes from the team at Orium. Smarter Patterns considers the form and function of the patterns as well as the ethical challenges that artificial intelligence poses. The focus lies on fostering transparency, trust, autonomy, fairness, and inclusiveness. Two wonderful resources to help us get our products fit for AI while putting the needs of people and the community first. (cm)
2. AI Under The Hood
There’s a lot of noise around language models like ChatGPT and their impact, but not enough clear-heading guidance about how to think about them or work with them. Maggie Appleton wants to change that. At SmashingConf Freiburg last year, she held a brilliant talk on the challenges of designing with language models. The video and transcript from the talk are available online. (Psst. Maggie will be live on stage at SmashingConf Antwerp this year, too!)
In “Squish Meets Structure: Designing with Language Models,” Maggie explores what language models are, how they came into existence, and why they are hard to design with. Maggie compares them to squishy, biological things that work in a completely different way to predictable software and the traditional programming logic we all are used to. The talk is a fantastic opportunity to reconsider existing mental models to better understand the behavior of AI and how to constrain it. (cm)
3. EU AI Act Cheat Sheet
Proposed by the European Union, the AI Act is the first comprehensive law on AI by a major regulator anywhere. It aims to regulate the use of AI, imposing extensive requirements on the development, deployment, and use of high-risk and general-purpose AI applications. The act is still in negotiation, but once effective, it will apply to organizations worldwide.
To give you a better idea of the scope of the AI Act, Oliver Patel published a handy cheat sheet that gets you familiar with the key elements of the act in just a minute. It will be updated once the full legislation text is published. As organizations have approximately 1–2 years to prepare to be compliant, the cheat sheet is a great opportunity to better assess the implications of the AI Act on your business already today. (cm)
4. Why Chatbots Aren’t The Future
Entering a prompt in a text box. That’s what AI tools mostly look like these days. But is this the future of interfaces? No, argues Amelia Wattenberger. In her post “Why Chatbots Are Not The Future,” she explores the problems of chat-based interfaces and where she sees the future of AI instead.
According to Amelia, the real game-changers have very little to do with plain content generation and blindly improving efficiency as we know it from popular AI tools today. Instead, she sees the real value of AI in embracing our humanity and building tools that help us gain clarity in our thinking.
Tom Cleary is also skeptical about prompt-driven AI, particularly seen from a usability perspective. He wrote a great blog post in which he examines what he calls the “prompting gap.” It lies in the fact that generative AI relies on users telling the machine what they want to achieve, although they often don’t know what they are looking for or don’t have the literary skills to be able to articulate it through prompts. A thought-provoking read. (cm)
5. Upcoming Workshops and Conferences
That’s right! We run online workshops on frontend and design, be it accessibility, performance, or design patterns. In fact, we have a couple of workshops coming up soon, and we thought that, you know, you might want to join in as well.
As always, here’s a quick overview:
- Design System Planning and Process Workflow
with Nathan Curtis. April 18–26 - Typography Masterclass Design
with Elliot Jay Stocks. May 1–15 - Behavioral Design Workshop UX
with Susan and Guthrie Weinschenk. May 6–20 - Scalable CSS Masterclass Dev
with Andy Bell. May 9–23 - Accessibility for Designers UX
with Stéphanie Walter. June 3–12 - Smart Interface Design Patterns Video Course UX
9h-video + Live UX Training with Vitaly Friedman - Jump to all workshops →
6. AI And Design Systems
Which opportunities do AI tools offer for design systems? Brad Frost and the team at Big Medium are finding that AI tools can help supercharge design system efforts across many categories, assisting in various aspects of creation and consumption. In his blog post “AI and Design Systems,” Brad demonstrates possible use cases.
As Brad shows, AI can be useful for writing component code — he compares it to a junior developer contributing code for review. AI tools also come in handy when migrating between tech stacks to save you time and reduce errors. By running unit tests and accessibility reviews, AI can help with testing and QA as developers focus on shipping the project. And AI can also take care of one of the most unpleasant tasks around design systems: documentation.
To ensure AI enhances a team’s work and doesn’t replace their jobs, Brad advocates for a human-centric mindset and a healthy level of skepticism when using the new technology. The guiding principles and considerations he shares in the post are a good starting point if you plan to incorporate AI tools into your workflow. (cm)
7. How To Integrate AI In Design
AI offers promising possibilities, yet we are still at the very beginning of learning how to integrate it into a design in a meaningful way. So, how do other organizations make use of the new opportunities? What’s their vision for enhancing their products with AI? The customer insights hub Dovetail shares a look inside their AI vision and the AI-powered features they are working on.
The focus of Dovetail’s AI vision lies on speed, accuracy, and discovery to enable users to gain the customer insights they need more quickly. The great potential of using AI also lies in reducing bias when analyzing insights and protecting customers’ privacy when sharing their data with stakeholders. A good example of what a roadmap for using AI in a way that supports its users and complements and augments the way they work can look like. (cm)
8. Generative Search
Imagine you plan to purchase a new chair for your home office. How do you decide which one to get? You’ll probably start googling something like “ergonomic office chair,” and then you go down the rabbit hole, combing through product descriptions, reviews, and comments to make your decision. The days of doing search like this might soon be over, though, as Google is experimenting with a Generative Search Experience.
Instead of just returning links, Google’s Generative Search Experience uses AI to generate an organized presentation of the subject, with a brief summary, bullet-pointed lists, and several jumping-off points to do more digging.
Daley Wilhelm wrote a great post in which she dives deeper into the Generative Search Experience, how it could change the user journey, and what this means for sites that generate income from ads. An interesting outlook at where things might be heading with AI — and a good reminder to do your own research from reputable sources before you make a decision on what to buy, where to have dinner, or who to vote for. (cm)
9. News From The Smashing Library 📚
Promoting best practices and providing you with practical tips to master your daily coding and design challenges has always been at the core of everything we do at Smashing.
In the past few years, we were very lucky to have worked together with some talented, caring people from the web community to publish their wealth of experience as printed books. Have you checked them out already?
- Understanding Privacy by Heather Burns
- Touch Design for Mobile Interfaces by Steven Hoober
- Image Optimization by Addy Osmani
- Check out all books →
That’s All, Folks!
Thank you so much for reading and for your support in helping us keep the web dev and design community strong with our newsletter. See you next time!
This newsletter issue was written and edited by Geoff Graham (gg), Cosima Mielke (cm), Vitaly Friedman (vf), and Iris Lješnjanin (il).
Smashing Newsletter
Front-end, design and UX. Sent out 1× a week.You can always unsubscribe with just one click.