Sunday 06 April 2025
As generative AI image tools continue to evolve, researchers are focusing on improving their usability and flexibility for designers and creatives. A recent study explored how users prefer to interact with these tools, shedding light on the various methods they employ to guide the AI’s output.
The study involved seven professional designers who were tasked with refining images using three primary input methods: text prompts, annotations, and scribbles. The researchers observed that each method had its strengths and weaknesses, depending on the specific task at hand.
Text prompts, for instance, proved effective when participants needed to make broad changes to an image’s overall mood or atmosphere. By providing detailed descriptions of what they wanted to achieve, designers could rely on the AI to fill in the gaps and create a cohesive visual representation. However, text prompts also had limitations, particularly when it came to specifying precise details or making minute adjustments.
Annotations, on the other hand, were often used to pinpoint specific areas of an image that required modification. Participants employed arrows, circles, and other symbols to draw attention to particular elements, such as a character’s facial expression or the placement of objects within a scene. Annotations allowed designers to provide more nuanced feedback than text prompts alone, but they still relied on the AI’s interpretation to accurately implement their intended changes.
Scribbles, meanwhile, were used to convey a sense of rough design or intent, often in combination with annotations and text prompts. By sketching out the general shape and form of an object or character, participants could provide a more organic and intuitive input for the AI to work from. Scribbles also enabled designers to explore different creative possibilities without being too prescriptive.
The study highlights the importance of flexibility and adaptability in generative AI image tools. By offering multiple input methods and allowing users to combine them in various ways, these tools can better support the diverse needs and workflows of designers and creatives. As the technology continues to advance, it will be crucial for developers to prioritize user experience and usability, ensuring that their products remain accessible and effective for a wide range of applications.
The findings also underscore the need for more advanced AI capabilities, particularly in terms of understanding and interpreting user input. As designers become increasingly reliant on these tools, they require systems that can accurately capture their intent and produce high-quality results with minimal effort.
Cite this article: “Unlocking Human Creativity: How Generative AI Image Tools Respond to Visual and Textual Prompts”, The Science Archive, 2025.
Ai Image Tools, Generative Ai, Usability, Flexibility, Designers, Creatives, Text Prompts, Annotations, Scribbles, User Experience, Workflow.
