In human-centered AI, UX and software roles are evolving

Were you not able to attend the event? You can find all of the summit sessions in our library. Here is the place to watch.

Two types of experts are needed for software development. There are people interested in how users interact with applications. The code that makes it work is written by those people. There is a boundary between the user experience designer and the software engineer. The advent of human-centered artificial intelligence is a challenge.

User interface designers use their understanding of human behavior and Usability principles to design graphical user interface. Hariharan Subramonyam is a research professor at theStanford Graduate School of Education and a faculty fellow of theSTANFORD Institute for Human-Centered Artificial Intelligence.

Subramonyam and three colleagues from the University of Michigan have developed recommendations for how the two can communicate in the age of artificial intelligence. There are practical steps and documentation that the two disciplines can use to convey the low-level details of their vision.

Read the study: Human-AI Guidelines in Practice: The Power of Leaky Abstractions in Cross-Disciplinary Teams

According to Subramonyam, the disciplines leak key information back and forth across what was once an impermeable boundary.

MetaBeat is a sequel to Meta Beat.

Thought leaders will give guidance on how metaverse technology will transform the way all industries communicate and do business in San Francisco on October 4.

Less is not always more

Subramonyam points to facial recognition as an example of a challenge presented by artificial intelligence. The unlocked interface used to be easy to describe. The user is using their fingers. The key pad is visible. The password is entered by the user. The application is authentic. The phone is accessible to the user.

User experience design begins to go deeper than the interface into the artificial intelligence. It is important for designers to think about things they have never thought of before. To describe how things should work in an ideal world is hard for designers. Engineers are finding that they can't build software to specific specifications. Training data can be considered a non-technical specification. Someone else is responsible for training data.

Engineers and designers have different priorities and incentives, which makes it hard for them to work together. Leaky abstractions are helping to reduce the amount of tension.

Radical reinvention

To understand how professional collaborations are evolving to meet the challenges of artificial intelligence, Subramonyam and colleagues interviewed 21 application design professionals.

There are a number of leaks for software engineers and designers to share information. The sharing of qualitative codebooks to communicate user needs in the annotations of training data is one suggestion for the UX designers. Optimal user interactions can be storyboarded by designers. They could record user testing to give examples of bad behavior. They suggest that engineers be invited to participate in user testing.

For engineers, the co-authors recommended the creation of computational notebooks of data characteristics, as well as visual dashboards that establish artificial intelligence and end- user performance expectations.

The main recommendation is for collaborating parties to delay committing to design specifications. The disciplines have to fit together. It's an easier fit if there's less complexity. It takes a while to get those edges polished.

There is sometimes a mismatch of needs in software development. If I, the engineer, create an initial version of my puzzle piece and you, the UX designer, create yours, we can work together to address misalignment over multiple iteration. Do we make the application specifications final at the last moment?

The historic boundary between engineer and designer is the enemy of good human-centered design according to Subramonyam.

Andrew is a contributor to the institute.

The story was first published on Hai.stanford.edu It's called Copyright 2022.

The VentureBeat community welcomes you.

Data decision makers can share data related insights and innovation.

Join us at DataDecisionMakers to read about cutting-edge ideas and up-to-date information.

You could possibly contribute an article of your own.

Data decision makers have more to say.