Building an ethics tool inside the tab people already use

The fastest way to kill an ethics tool is to make it one more app people need to open. So I built AI Ethics Coach as a Chrome extension that runs where the writing already happens.

What I built

I built an MV3 extension with a selection flow, quick review surface, and deeper panel view for follow-up. The system sends selected text for analysis and returns issue flags users can act on.

Why I built it

I wanted something that helps with practical review decisions instead of sounding like a lecture.

The build

• MV3 extension scaffolding
• selection-triggered review workflow
• UI split across quick and detailed views
• iterative wording cleanup to keep outputs readable

What broke and what I learned

The early UI over-explained every warning. That made users skip the content. I learned that shorter, concrete flags beat long moral commentary.

Current status and next step

Current status is a functional extension story with a clear interface pattern. Next step is collecting better real screenshots for public posting and improving issue grouping.

Links:

• GitHub profile: https://github.com/pradyotbathuri
• Portfolio case: portfolio/ai-ethics-coach-extension.md