Conversational User Interface (CUI) is one of the most popular ways of interacting with AI today. The user types in a request using text and then receives a text-generated response. However, block of text is a very limited form of information consumption for interface users.
Share article
Conversational User Interface (CUI) is one of the most popular ways of interacting with AI today. The user types in a request using text and then receives a text-generated response. However, block of text is a very limited form of information consumption for interface users. Rethinking AI interfaces providing written content from the perspective of Information Design and Data Visualisation can unlock the potential that lies in the visual representation of information. What’s more, it can also help users to build mental models from AI-generated content more effectively.
Building prompts has already become a field of knowledge to achieve much better results in collaborations with AI. At the moment, dozens of prompt courses can be found to help users communicate better with AI. But is an interface that you first have to learn for several hours a truly good and user-friendly interface? On the one hand, a blank field and the opportunity to ask any question or command creates an abundance of possibilities; on the other hand, this excess becomes a problem for users when they do not know completely how to start. Also, after the first wave of excitement subsides, in the long term, using the interface and constantly typing would become tiring for users.
A good UX can help by designing the interface to build the visual context for the question or command and guide the user through the correct prompt creation. What’s more, this approach also avoids the frustration of poor quality AI output due to a poorly asked prompt. AI applications used to achieve specific goals, e.g. planning a visit to a particular city or country, taking into account weather conditions, preferred way of transport, place of accommodation, favourite cuisine, etc., will particularly benefit from the visual construction of the prompt.
Similarly, an interesting area for UX designers is the way AI delivers information to the user. From a design perspective, it represents a return to familiar areas such as Information Design and Data Visualization. It involves not only considering the interaction and how users engage with AI, but also how the information itself is structured and presented. A well-designed layout, coupled with proper information extraction, will guide users through the content. Additionally, highlighting the connections between different pieces of information and revealing their relationships can present it in a simpler and more comprehensive manner.
Scopus is an academic database that collects scientific papers from all disciplines, providing access to the worldwide academic community on a single platform. Typically, searches are conducted by entering phrases and keywords, after which the application provides the user with a list of scientific papers related to the query.
Recently, developers unveiled a new approach to scientific search called Scopus AI. The query is not only analyzed based on the occurring words, but the model can also better understand the overall context. Furthermore, the designers have redesigned the Scopus interface to enable users to uncover relationships between the provided information and explore further areas of interest interactively. The redesign of the user interface has enabled the provision of brief summaries, reference lists, and filtering by topic experts, among other possibilities.
What is more, the designers have also introduced a new feature called Concept Map, which allows content to be explored using an interactive map representing connections between scientific topics. The visual representation of information can allow users to place the topic they are researching into a broader context even more easily. The presented feature appears to be an interesting start, but in my opinion, it still holds enormous potential that, if explored by the designers, could truly revolutionize AI design. The entire presentation of the new Scopus AI interface can be found here.
In conclusion, enhancing AI interfaces with visual elements, guided prompt creation, and intuitive information structuring not only fosters more effective communication between users and AI but also facilitates the construction of mental models, ultimately revolutionizing the user experience and unlocking the full potential of designing for AI.
