TRANSPARENCY-CHECK
The instrument “Transparency-Check” offers guidance for the design and study of transparency in
Artificial Intelligence-based systems that use personalization techniques. Transparency is a practice
of system design that centers on the disclosure of information to users, whereas this information
should be understandable to the respective user and provide insights about the system. Specifically,
the information disclosed should enable the user to understand why and how the system may
produce or why and how it has produced a certain outcome. Transparency is increasingly important
for responsible computing, and designers are pressured to implement transparency due to ethics,
laws, and performance pressure.
To offer guidance, Transparency-Check was developed by a research team at the University of Tübingen in Germany and Ben-Gurion University in Israel. Responsible for Transparency-Check are Laura Schelenz, Dr. Avi Segal, and Prof. Kobi Gal.
1
Who can use Transparency-Check?
The instrument is free for everyone to use as long as they agree to the terms of use and privacy
policy (see below). Transparency-Check is developed for designers, researchers, and end users.
Designers of AI-based systems can use Transparency-Check to review the overall level of
transparency in their own system; assess in which areas (data collection, data processing, output,
and user control) their system is strongest regarding transparency; stimulate discussion in their
design team about the multi-faceted aspects of transparency.
Researchers can use Transparency-
Check to study and compare transparency levels of different systems in the wild; study user
perceptions of transparency in real-world or synthetic systems; develop design best practices for
implementing transparency.
2
What exactly is Transparency-Check?
Transparency-Check is a two-part tool of analysis that helps designers, researchers, and users
understand the levels of transparency in different systems. In the first part of Transparency-Check,
which consists of 7 prompting questions, the user of Transparency-Check can adapt the instrument
to their specific use case, context, or domain.
In the second part of Transparency-Check, a transparency checklist of
19 questions guides the user of the instrument through an evaluation of the system in question.Follow the link below to access Transparency-Check.
3
What do I need to get started?
You need to be interested in the transparency of a computer system that uses personalization. If you
are interested in several systems or want to compare the levels of transparency in different systems
or the levels of transparency for the same system over time, it is recommended to use Transparency-
Check several times. There is no limitation about the number of times you can use the instrument.
When you start using Transparency-Check, please have access to the system you review as well as
the website, links, or other services tied to the system, as you might have to research some
information to be able to answer the questions.