Obviously there’s a learning curve but a “set of rules for x purpose” would be so nice.

The Challenges of Custom Card Design with AI Assistance: The Need for Clear Guidelines

Creating custom digital assets can be a rewarding yet challenging process, especially when relying on AI tools like ChatGPT for assistance. While AI can streamline certain aspects of design, many users encounter a steep learning curve, often wishing for a comprehensive set of best practices or guidelines tailored to specific tasks.

Navigating the Learning Curve in AI-Driven Design

As a user working on designing custom game cards, I initially found ChatGPT to be quite helpful. Generating initial card fronts with minimal fuss and making straightforward edits were smooth processes. However, complexities arose when moving to the card backs, revealing inherent limitations in the current workflow and prompting questions about the best practices.

Common Hurdles Encountered

One recurring issue involved the AI inadvertently transferring symbols or elements from the front side to the back—unintended overlaps that disrupted the design. Attempts to simplify the process by removing auxiliary tools, such as a custom “toolkit” used for the front design, temporarily helped but didn’t fully resolve the problems.

Further complications emerged when trying to finalize the backside. For instance, unexpected artifacts—like “tongues” of leftover characters—appeared on the design, resembling tangled spaghetti rather than clean text. Clarifying instructions to exclude specific elements proved insufficient because ChatGPT struggled with nuanced visual editing tasks.

Limitations of AI-Based Editing

Using AI to clean up images or remove unwanted sections often led to overcorrections or residual artifacts. For example, highlighting areas for removal and instructing the AI to delete them frequently resulted in entire sections being erased, sometimes leaving behind remnants of color or texture that needed additional manual correction.

Additionally, changing design elements such as highlighters or background colors involved identifying specific color codes (hex or RGBA). However, removing these color overlays from baked images was often complicated, requiring external tools or extensive trial-and-error.

The Need for Structured Guidelines

Throughout this process, it became evident that many issues could be avoided if there were clear, accessible guidelines or “rules of thumb” for each task—akin to a comprehensive manual for custom card creation. Knowing the best sequence of steps, what to avoid, and how to troubleshoot common pitfalls would significantly streamline the workflow, saving time and frustration.

Conclusion: Advocating for Standardized Best Practices

While AI tools are powerful, their effectiveness depends heavily on the user’s understanding of optimal workflows and best practices. The absence of clear