This diploma thesis analyzed the process of designing modern packaging using generative artificial intelligence (GenAI) and traditional design approaches. The research aimed to determine how GenAI influences the design process and to evaluate its role in comparison to conventional methods.
In the practical part, a packaging concept for a thermo tumbler was developed. The process included an analysis of the product, target group and existing packaging, followed by the creation of user personas, use cases and information architecture. Various tools were used, including the Magnific platform for generating images, ChatGPT for creating structured prompts, and Adobe Photoshop, Adobe Illustrator and Adobe Dimensions for further processing and final design development. Visual outputs generated by different GenAI models (GPT 1.5 High, Google Nano Banano 2, Flux.2 Pro and Seedream 5 Lite) were systematically compared based on criteria such as realism, preservation of the product's composition, and consistency with the reference image.
The results showed that GenAI enables rapid generation of visual solutions and efficient exploration of different design directions. However, limitations were identified, especially in more complex scenes, where errors such as deformations, incorrect composition and unwanted elements occurred.
It was concluded that GenAI functions as a supportive tool that requires critical evaluation by the designer. The combination of generative and traditional design approaches proved to be the most effective, as it balances speed, creativity and control over the final solution.
|
