Salvato in:
Dettagli Bibliografici
Autore principale: Khor, Ashley
Natura: Recurso digital
Lingua:inglese
Pubblicazione: Zenodo 2025
Soggetti:
Accesso online:https://doi.org/10.5281/zenodo.15570677
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866901197328220160
author Khor, Ashley
author_facet Khor, Ashley
contents <p>This archive contains the presentation materials for <strong>“Beyond Compliance: Generative AI Safety Evaluation for Civil Society,”</strong> delivered by Ashley Khor at the IHS Summer Graduate Conference on May 31, 2025. The talk introduces a hybrid evaluation framework designed specifically for high-impact, community-driven AI use cases. Developed through both academic review and field practice, the framework incorporates five core dimensions of safety:</p> <ol> <li> <p><strong>Technical and content integrity</strong></p> </li> <li> <p><strong>Relational and ethical grounding</strong> (design justice, feminist, and trauma-informed principles)</p> </li> <li> <p><strong>Usability and accessibility</strong></p> </li> <li> <p><strong>Organizational readiness</strong></p> </li> <li> <p><strong>Contextual appropriateness</strong> for sensitive, user-directed interactions.</p> </li> </ol> <p>In addition to defining these dimensions, the project extends current GenAI safety practices through a participatory methodology that includes five test-a-thon methods: <strong>human-in-the-loop scoring</strong>, <strong>participatory red teaming</strong>, <strong>simulated user journeys</strong>, <strong>reflective feedback circles</strong>, and <strong>observer-based shadow scoring</strong>. Together, these methods offer a survivor-centered and emotionally attuned lens for evaluating AI systems — not only by their outputs, but by how they are experienced, trusted, and used.</p>
format Recurso digital
id zenodo_https___doi_org_10_5281_zenodo_15570677
institution Zenodo
language eng
publishDate 2025
publisher Zenodo
record_format zenodo
spellingShingle Beyond Compliance: Generative AI Safety Evaluation for Civil Society
Khor, Ashley
Artificial Intelligence
Artificial Intelligence/standards
Artificial Intelligence/ethics
Community-Based Participatory Research/methods
Design Justice
Survivor-AI
Generative AI Safety
Trauma-Informed Computing
LLM Evaluation
Public Interest Technology
AI Trust & Safety
<p>This archive contains the presentation materials for <strong>“Beyond Compliance: Generative AI Safety Evaluation for Civil Society,”</strong> delivered by Ashley Khor at the IHS Summer Graduate Conference on May 31, 2025. The talk introduces a hybrid evaluation framework designed specifically for high-impact, community-driven AI use cases. Developed through both academic review and field practice, the framework incorporates five core dimensions of safety:</p> <ol> <li> <p><strong>Technical and content integrity</strong></p> </li> <li> <p><strong>Relational and ethical grounding</strong> (design justice, feminist, and trauma-informed principles)</p> </li> <li> <p><strong>Usability and accessibility</strong></p> </li> <li> <p><strong>Organizational readiness</strong></p> </li> <li> <p><strong>Contextual appropriateness</strong> for sensitive, user-directed interactions.</p> </li> </ol> <p>In addition to defining these dimensions, the project extends current GenAI safety practices through a participatory methodology that includes five test-a-thon methods: <strong>human-in-the-loop scoring</strong>, <strong>participatory red teaming</strong>, <strong>simulated user journeys</strong>, <strong>reflective feedback circles</strong>, and <strong>observer-based shadow scoring</strong>. Together, these methods offer a survivor-centered and emotionally attuned lens for evaluating AI systems — not only by their outputs, but by how they are experienced, trusted, and used.</p>
title Beyond Compliance: Generative AI Safety Evaluation for Civil Society
topic Artificial Intelligence
Artificial Intelligence/standards
Artificial Intelligence/ethics
Community-Based Participatory Research/methods
Design Justice
Survivor-AI
Generative AI Safety
Trauma-Informed Computing
LLM Evaluation
Public Interest Technology
AI Trust & Safety
url https://doi.org/10.5281/zenodo.15570677