For some time now, there has been talk of the possible applications of artificial intelligence in various work sectors. Attention seems to be focused mainly on those complex forms of generative AI capable of generating content very similar to what a human being could provide, starting from simple user requests. The UX sector also appears to be affected by these means, and in the following article, we will specifically evaluate the impact that ChatGPT and Google Bard are having in this context. In this regard, we will consider the advantages that their use can bring, the main restrictions they have, and some practical examples.
What are Google Bard and ChatGPT?
Developed by Google AI and OpenAI, Google Bard and ChatGPT are both AI-based chatbots. They are examples of generative AI technology, capable of processing natural language and understanding aspects such as grammar and text structure. In this way, they can provide precise and coherent responses to user requests through a chat interface.
Launched for the first time on November 30, 2022, ChatGPT (GPT is an acronym for Generative Pre-Trained Transformer) is a language model initially created through supervised training techniques. These techniques involve providing the AI system with examples of input and output in the form of simulated conversations, where trainers play the roles of both the machine and the user. By analyzing the accumulated data, the machine learning model detects recurring patterns, enabling it to anticipate responses to novel inquiries. In addition to this process, reinforcement training was employed. In this case, trainers classified various responses based on their quality, and the machine learned to provide solutions that maximize its reward.
Similarly, Google Bard is a chatbot that, based on information obtained from Google search results, can provide the latest information available on the search engine. First published on January 18, 2023, exclusively in England and the United States, it has been made available in many more countries, including Italy, starting from July 13, 2023, in over 40 different languages.
The main differences between the two chatbots lie mainly in the language models they use. OpenAI’s artificial intelligence utilizes the GPT 3.5 model in its free version (trained on data updated until 2021) and GPT4 for the paid version; Bard, on the other hand, is based on the more recent PaLM-2. Additionally, while Google’s AI provides more blocks of information in its responses, ChatGPT generates a single coherent text. Both ChatGPT and Google Bard are nevertheless large language models (LLM), capable of generating textual content of various types based on acquired knowledge, similar to responses that a human could provide. Among the most frequently cited use cases are:
- Generating programming code or parts of it in different languages and instructions;
- Writing emails, posts, and messages of various types;
- Providing explanations and in-depth information on various topics;
- Producing creative content such as poetry, stories, songs, and scripts.
What opportunities does AI offer in user experience design?
If we consider the speed of processing, good usability, and the widespread accessibility of these tools, we can recognize them as collaborative and adaptable means for different contexts and requirements. In the same way that graphic design programs assist graphic designers, the use of AI-based chatbots can positively impact the work of UX designers and researchers.
Engaging in user experience design means being part of a constant learning process, highlighting and refining the key aspects that guide the user during their interactions with digital products and services. The primary focus is on emotions, experience, and the feelings users have during the navigation of a website or the use of mobile apps. The goal of this process is oriented towards ensuring practical, enjoyable, and intuitive use, identifying aspects to address, and optimizing them.
In this sense, ChatGPT and Google Bard can be tools that facilitate the work of UX designers and researchers, making their work more efficient by exploring different design ideas, perhaps overlooked initially, and optimizing time spent on more mechanical and less creative tasks. In this regard, a recent empirical study conducted by MIT members showed how the use of ChatGPT by professionals in different business sectors for drafting professional documents reduced working time, and also resulted in higher-quality texts.
The Use of AI for UX Research
As programs with access to vast amounts of data sets, Bard and ChatGPT can offer valuable information for conducting research, providing, for example, demographic data on a specific target of potential users. Another example is the ability of these programs to generate prototypes of user personas, fictional but realistic representations of audience segments to address in the design of products and services.
AI application in UX Design
Although these programs are currently only capable of producing textual outputs, their use in UX Design should not be underestimated; indeed, an exploratory use of AI can assist in generating ideas and inspirations that can be given a visual representation. Clear examples include the ability to generate representations of user flows for different types of applications or websites or the ability to identify major usability issues in the use of products and services.
The Use of AI in UI Design
The use of Bard and ChatGPT can extend to generating ideas and inspirations related to user interface design. In this case, they can be used to create web pages for inspiration, guiding how a landing page for a retail website might look, the tone it should convey, logo design, or the best colors to use to communicate certain emotions to users.
AI for Copywriting
Copywriting is a clear example of how the application of chatbots can be valuable in user experience design. These language models can offer ideas for drafting notifications, error messages, onboarding, and informational texts, optimized based on the designer’s needs, making them more or less concise, creative, and conveying the preferred tone of voice.
To summarize, this work translates into a form of Generative Design, a model for creating possible solutions based on criteria established by designers, setting the limits and goals they should meet. The proper usability of these tools, along with their speed in responding, allows for the generation of rough prototypes, useful for exploring design options that might not be considered.
Limitations of Artificial Intelligence for UX Design
Generative artificial intelligence tools can play a role in user experience design. In particular, more “mechanical” and iterative tasks, which often consume time during the creation phases, can be simplified through these means. However, concerning the more creative aspects of design, these tools still seem incapable of replacing people; their use must necessarily be supervised and constantly enriched through the use of data and analytics by the designers who employ them. We can consider some of the more limiting aspects that distinguish tools like Bard and ChatGPT.
- Artificial intelligence systems operate thanks to the data and patterns they have been trained on. The datasets used are never “neutral” but always reflect the biases and prejudices of the people who produced them. For this reason, such tools risk amplifying these biases, making the responses valid with specific languages and ways of thinking but not legitimate for others.
- Designing the user experience of a digital product or service means understanding the needs and interests of the people for whom they are developed. Therefore, it is essential to highlight the aspects that make the use or navigation of a platform more practical and enjoyable, while identifying the major issues that may arise in its use. Artificial intelligence cannot empathize and show sensitivity, and therefore, they cannot understand the psychological and emotional needs that guide people. In the development of a fully engaging experience, it is necessary to take into account the shortcomings of these language models in this regard.
- Bard and ChatGPT intervene in the exploratory phase, providing ideas and insights to explore alternative solutions or consider problems from different perspectives. However, using these tools on a large scale also means giving rise to content that may be similar to each other, leading to a lack of originality and personalization. If the training datasets of language models are the same, the solutions offered will also be similar and may not meet the individual needs of individual users.
- It is also necessary to consider that Bard and ChatGPT are not exempt from errors and inaccuracies. Susceptibility to how requests are expressed by users and the specific terms used can lead them to respond in different ways, refining responses but also stating that they do not know the solutions to problems. OpenAI itself acknowledges in this regard how its language model is capable of providing plausible but incorrect or nonsensical answers.
Examples of Optimization with ChatGPT and Bard for UX Design
Having access to such complex language models allows us to incorporate them into the design process in various ways. While the designer’s role is irreplaceable for more creative tasks, many iterative processes can be automated through their use. The use of clear and increasingly refined prompts helps obtain responses and ideas suitable for the problems at hand. The particular usefulness in this regard is that Bard and ChatGPT recognize the entire conversation with the user and can refine responses based on how the prompts are formulated.
We can observe some examples where the use of chatbots can be helpful for UX/UI designers.
- Generation of User Personas: A key step in designing a product or service is understanding the target audience. In a human-centered design, it is essential to build personas that represent this audience. By including information about the type of product being conceived in the instructions to artificial intelligence, it is possible to obtain user persona models useful for guiding the design process.
Example prompt: Create a user persona for a user of a dating website.
- Design System: While they cannot offer graphic representations, ChatGPT and Google Bard can be used to improve design systems and come up with design ideas, taking into account information related to the type of product or user.
Example prompt: Provide an example of a design system for a button. Represent it in a table with various states and design tokens.
- Competitor Analysis and Comparisons: ChatGPT and Bard offer the opportunity to assist designers in comparing different applications and websites, quickly identifying content and features.
Example prompt: Conduct a comparison between these food delivery sites [links to the sites].
- Generation of Texts and Suggestions for Structuring User Research: Examples include generating ideas to structure questions for interviews and surveys to research users of digital products and services.
Example prompt: Write 10 questions for a survey describing the use of a travel website.
- Data Synthesis: Bard and ChatGPT can interpret long lines of text, identifying elements such as keywords or more relevant information.
Example prompt: Summarize the content of the following [text], indicating keywords.
- Ideas in Copywriting: Examples of this use may include creating messages in the user interface or providing insights for writing articles, posts, and emails.
Example prompt: Write an introductory email for an interview with a user of an application.
These are just a few examples of how these types of artificial intelligence can impact the work of UX designers. Learning to incorporate these powerful tools into our work and refining the instructions given to them can open the doors to new possibilities of use.
In this article, we have considered a specific kind of generative AI, the one of chatbots capable of generating textual content, highlighting the opportunities and limits that can transform the research and design work of user experience. Certainly, the potential promised by these means can go far beyond, changing depending on how designers and researchers refine their skills in their use. In this sense, asking whether artificial intelligence can automate the work of UX professionals calls for a complex answer: different tasks can now be effectively performed by these machines, but the added value represented by the expertise, capabilities, and creativity of experts will always be valuable and essential.
Therefore, we are moving in a context where we have new capabilities and powers. However, the speed with which we can design and invent new solutions in the design of products and services calls for new responsibilities to which we are called to respond. Nowadays, more than ever, we can design a user experience supported by vast amounts of data and innovative means with which we can collect and process them.
In this sense, we have the opportunity to change our perspective, bringing together both a microscopic view, focused on individual user interactions, and a macroscopic perspective, addressing larger issues, relationships with stakeholders, new scenarios, and competing objectives. New tools, technical instruments, and procedures can help make user experience research faster and more intuitive, allowing for constant and proactive data collection that does not only begin and end with the onset of specific problems.
The challenge for companies like Conflux is therefore to use new tools not only to make the UX design work easier and faster but especially to support this process through constant data and information collection from which to learn more and more. This is the maximum added value that humans can bring into a real world increasingly influenced by artificial intelligence.
If you are looking for a partner to create a successful digital product or online service, Conflux is the perfect UX agency for you. Through the combination of our professionals’ expertise and the use of innovative tools, we will be able to ensure a tailored user experience for you and your customers.