An emergency-shelter intake team’s retro comments, when summarized by AI, systematically downplay overnight-staff concerns. What should the Team Coach flag?
Select an answer to reveal the explanation.
Short Explanation
If the summary always “forgets” the overnight shelter shift, that is bias wearing a neat bullet list. Read the real comments before anyone acts on the digest. Sensitive team feedback still needs a human pair of eyes.
Full Explanation
When AI summaries of an emergency-shelter intake team's retrospective systematically downplay overnight-staff concerns, the Team Coach should flag that AI summaries can encode bias and keep human review of sensitive team feedback. Overnight concerns are operationally material for shelter safety and staffing. A neat bullet list that "forgets" a shift is bias, not neutral compression.
Dropping overnight-staff comments because the model ranked them low importance fails because model ranking is not a substitute for operational judgment about who must be heard. Publishing the biased summary to the city board as the official retrospective record fails because it launders skewed feedback into public governance. Replacing human retrospectives with the model so shelter staff no longer speak fails because it silences the people closest to overnight reality.
Exam caveat: summarization aids are useful only when humans can still see and correct whose voices were under-weighted. Operational check: read raw overnight comments beside any AI digest, invite those staff into the improvement conversation, correct or discard biased summaries before external sharing, and keep the human retrospective intact.