From Risk Avoidance to User Empowerment: Reframing Safety in Generative AI for Mental Health Crises
arXiv 2603.05647•2cc30cdebb59c8cdd5239eabba85fc5b34cc4540eced69515b57afe2fb784139
AI researcher surveyLLM risksLoRATikTokadvertising lawalgorithmic auditingambiguity collapseaugmentative and alternative communicationbias mitigationbiometricscomputational social sciencecontent moderationcrisis supportdomain-adaptive fine-tuningfairness in hiringgenerative AIhuman-in-the-loopinfluencer marketingliability avoidancemental healthminors protectiontopic modelinguser empowerment
Paper metadata
- arXiv ID
- 2603.05647
- Version
- Not specified by this published record
- Category
- Computer Science — Computers and Society (cs.CY)
The PDF link points to arxiv.org. Baitaphish does not expose a private stored PDF.
Evidence and limitations
- Source ID
- arxiv_cs_cy
- Record identifier
- 2cc30cdebb59c8cdd5239eabba85fc5b34cc4540eced69515b57afe2fb784139
- Enrichment time
- 2026-03-09T07:23:52Z
- AI-assisted enrichment
- Yes
This record may overlap with other records. Its enrichment can be incomplete or wrong, and machine assistance was used. Validate consequential decisions against the linked source and your own environment.