Get Full Government Meeting Transcripts, Videos, & Alerts Forever!
Get email alerts on the Deepfakes And Content Provenance topic
No spam. Unsubscribe anytime.
Subcommittee recommends provenance labeling and flags two Guam bills on deepfakes
Summary
The ethics subcommittee recommended focusing on provenance requirements (C2PA/SynthID) and penalizing credential stripping rather than rushing universal technology mandates; it identified two Guam bills under review addressing nonconsensual AI intimate images and deceptive AI election media.
The AI ethics subcommittee reviewed content-provenance technologies and local legislative activity addressing deepfakes and algorithmic exploitation.
The presenter described C2PA (Coalition for Content Provenance and Authenticity) as an emerging industry standard that cryptographically attaches origin credentials to media files and noted SynthID as a supplemental technology. The group acknowledged limits—credentials can be stripped and screenshots can defeat some protections—and recommended focusing policy on provenance requirements and penalties for bad actors who strip or misuse provenance rather than mandating a single vendor technology.
The subcommittee identified two Guam bills for focused review: one described in the presentation as "bill 171-38 prohibiting nonconsensual intimate AI depictions" and another as "bill 209 prohibiting materially deceptive AI media in elections." Both bills have had public hearings but had not advanced at the time of reporting. The subcommittee plans to draft three policy briefs on top harms and recommendations to the governor ahead of the statutory deadline.
Why it matters: provenance and labeling policy intersects privacy, sexual-harm protections, and election integrity. The subcommittee favored a rolling, actionable approach—releasing briefs as topics mature so recommendations can move forward without waiting for complete statutory work.
