Are there any AI tools that can validate design approaches?

No. The feedback itself is too complex for an AI to validate it: you're going to have to talk to a real human if you want good feedback.

ChatGPT can be pretty good at evaluating a first pass of your system design. It can give you a good first step about how to approach a problem as well as coming up with paths to go down for your specific use case. Of course, you need to be able to think about the solution as it relates to your company's architecture.

All of the AI chat bots (ChatGPT, Gemini, etc) are relatively decent-ish at this. However, you need to be specific with your prompt. In particular, you should lay out your system design values (e.g. velocity vs. quality, fail fast vs. suppress, what you value the most among resilience, scalability, correctness, etc).

My advice is to treat the AI as a friendly partner to bounce ideas off of and listen to yourself in a productive way, not as an authoritative source. This is because there is no such thing as a "right" system design due to system design being all about trade-offs and what your organization's specific engineering culture values. AI is much better suited for situations where there is a clear cut optimal solution like data structures and algorithms problems. With open-ended problem spaces, AI struggles immensely.

On top of all this, you should never have AI be your sole source of system design feedback. You should always run a technical design proposal by actual humans. A competent Staff Engineer will give you 10x better feedback on your design compared to AI, even if you gave the AI the most specific prompt in the universe.

No matter what you do, you need to make a detailed system design doc to either feed into the AI, your team, or both. Go through this course to learn how to do that: System Design Masterclass: Shipping Real Features To Production