Accessibility, Generative UI and Vivaldi (Re: Accessibility Has Failed: Try Generative UI = Individualized UX)

This is English translation of my previous post written in Japanese, so that anyone who speaks English can find and read it more easily a bit.

Jacob Nielsen is a well-known authority in the field of usability and UX, and the longer you have been in charge of web designing/development, the more familiar his name will be with you. For a period of time, I also studied with his books. The recent article "Accessibility Has Failed: Try Generative UI = Individualized UX" written by him is something I cannot overlook from the standpoint of a person who has been involved in the digital accessibility for many years.

Of course, I acknowledge that digital accessibility is still a ways off, but even in my narrow observation range, the article has been unpopular and is quickly starting to be challenged (e.g., On Nielsen's ideas about generative UI for resolving accessibility). By the way, here is my knee-jerk reply to the article I left on Mastodon and LinkedIn:

The idea of a generative UI is interesting and full of potential, but there should be no need to discredit past efforts in the accessibility industry to promote it. In addition, there is no need to portray the relationship between accessibility and generative UI as if it were a dichotomy.

This article vividly illustrates that even the biggest names in a field can undermine areas they are not familiar with in order to strengthen their own argument.

While I don't like the article as a whole, the idea of generative UI is interesting, and I see potential in it. I have also written a bit similar articles like "the direction in which web browsers should evolve is the ease of customizing the displayed content (in Japanese)" in the past. More recently, in an interview article with Vivaldi's founder, Mr. Tetzchner, titled "The Essence of Web Accessibility thinking with the Web Browser in the AI Era (in Japanese)", we had the following talk:

Kazuhito: At the moment, users can still adjust font size, line height, and color contrast of web content. I personally believe that further customization and personalization enhancements on the browser could be a game changer.

I understand that users can change the way how web content rendered in the browser technically, rather than display the content as it is intended by the author.

From such perspective, an ideal browser to me would be one that automatically generates user-friendly user style sheets and user scripts automatically. Currently, Vivaldi is the closest browser to that ideal. What do you think about my idea?

Tetzchner: Vivaldi allows fine-tuned customization of the browser's UI placement and size, so similarly, web pages should be free to change to meet users' needs.

Vivaldi has Page Action feature that allows you to change the way a web page looks, but does not yet have the ability to edit and share your own Page Actions.

Anyone with knowledge of CSS and JavaScript can create Page Actions that apply different looks and feels. In the future, we hope to have a place to share Page Actions similar to Vivaldi's theme gallery, so that everyone can use the web page the way it looks and feels right to them.

Kazuhito: Thank you. Let me share one more idea.

If Vivaldi can memorize results of customization/personalization of, for example, font size, line spacing, color contrast etc., while respecting author's intended visual design to some extent, the site will always be easy to view and use when accessing the same site. Furthermore, Vivaldi could use AI technologies to learn such customization/personalization trends and suggest settings that are easy to view and use in advance according to user characteristics, even for first-time visitors to a site. Is such a direction possible as one direction in which web browsers will evolve?

Tetzchner: There are privacy concerns regarding the part where AI automatically collects data on user actions, operations, page views, etc.

On the other hand, we believe that one good use of AI technology is to suggest how to configure based on users' needs, such as "minimize font size" or "read the screen out loud," or to output a style sheet.

Therefore, I find the idea of a generative UI appealing, and I am a proponent of the idea that web browsers should implement similar functionality. For that reason, I find the article's dismissal of all the progress that has been made in digital accessibility and the efforts of those who have gone before us to make that progress a reality to be extremely disappointing.

[ 2024-03-10 Update ] This post has been introduced at Jakob Has Jumped the Shark -- Adrian Roselli.