I am closing this as Resolved. Here's a summary of the results — I also shared a deck with @AAlhazwani-WMF and @KStoller-WMF.

Summary of results

Open exploration of the Dolphin article — first impressions & editing priorities

• Tone issues not naturally noticed
  • Several participants did not identify “great” as problematic, instead focused on factual accuracy or completeness (checking details, suggesting images)
  • Only one user explicitly identified “great” as unneutral.
  • Other priorities surfaced first: fact-checking, readability, media (images)... Tone may not be a newcomer’s first instinct.

Understanding ‘Improve tone’

• Interpretations varied:
  • General task understood: Most participants grasped the overall idea of suggested edits, but many did not focus specifically on tone as the editing goal.
  • Descriptions/interpretations mentioned:
    • Remove promotional tone
    • Make text simpler and factual

Tutorial experience

• All participants went through the tutorial
• Some users did not realize it could be skipped and only noticed the skip option later
• Positive reception: described as intuitive, short, very useful, and improved preparedness for tone revision.
• One user mistook the tutorial for the real, actual task. Another user only realized at the end that it was a tutorial.
  • Confusion started when they saw the ‘Edit’ button in S2, that action didn’t seem like it was going to launch a tutorial.

Editing after onboarding

• 3 users felt more confident identifying tone issues after the tutorial.
• Some users expressed mistrust in AI consistency and LLMs

Neutrality: views & definitions

• Mixed definitions and possible inconsistent understanding
  • 2 users described neutrality as clarity, conciseness, and avoiding exaggeration.
  • 2 users emphasized neutrality as avoiding promotional or opinionated language.
  • 1 user did not provide a clear definition.

I am available for questions if they come up!