Council reviews updated social media policy; staff add generative-AI standards and external-post guidance
Get AI-powered insights, summaries, and transcripts
Subscribe
Summary
City staff presented a revised social media policy that permits limited staff interaction on partner social channels during official business, adds emergency-misinformation language, and sets minimum human‑review and attribution standards for generative-AI content.
City Administrator Cynthia McNabb and staff presented a redlined update to Duval's social media policy, which adds three principal items: (1) a statement allowing city staff responsible for social media to "upon occasion" like, repost or respond to external partner social channels on city business; (2) a short emergency clause allowing the city to publish verified updates and limit social channels during an emergency or widespread misinformation; and (3) a new set of standards for generative artificial intelligence tools used in city social content.
The AI standards in the draft require that all AI-generated content "undergo a thorough human review for inaccuracies and validation before finalization, dissemination, or implementation," that AI content be treated as a starting point instead of a substitute for professional judgment, and that images or videos created exclusively by generative AI be clearly attributed. The draft also states that generative AI use must respect privacy and not amplify bias or stereotypes.
Council members and staff discussed whether interactions (likes/comments) by the city's social-account profile on external posts should be retained and treated the same as content posted on city-managed pages. Staff confirmed the city uses an archiving system (CivicPlus) that captures posts and metadata, and council asked legal staff to double-check retention and public-records implications. A council member asked staff to extend retention/disclosure language to cover "interactions made using the city's social media profile" (likes and comments on other pages); staff said platform and archiving tools already capture handles and interactions but agreed to review consistent language.
Council did not adopt the policy at the meeting; staff were asked to incorporate council edits, provide suggested wording about remote public comment procedures elsewhere in the procedures document, and seek legal review on retention, disclosure and whether to require explicit attribution for AI-generated images or video.