Evaluating the Social Impact of Generative AI Systems in Systems and Society
read the original abstract
Generative AI systems across modalities, ranging from text (including code), image, audio, and video, have broad social impacts, but there is no official standard for means of evaluating those impacts or for which impacts should be evaluated. In this paper, we present a guide that moves toward a standard approach in evaluating a base generative AI system for any modality in two overarching categories: what can be evaluated in a base system independent of context and what can be evaluated in a societal context. Importantly, this refers to base systems that have no predetermined application or deployment context, including a model itself, as well as system components, such as training data. Our framework for a base system defines seven categories of social impact: bias, stereotypes, and representational harms; cultural values and sensitive content; disparate performance; privacy and data protection; financial costs; environmental costs; and data and content moderation labor costs. Suggested methods for evaluation apply to listed generative modalities and analyses of the limitations of existing evaluations serve as a starting point for necessary investment in future evaluations. We offer five overarching categories for what can be evaluated in a broader societal context, each with its own subcategories: trustworthiness and autonomy; inequality, marginalization, and violence; concentration of authority; labor and creativity; and ecosystem and environment. Each subcategory includes recommendations for mitigating harm.
This paper has not been read by Pith yet.
Forward citations
Cited by 10 Pith papers
-
Debiasing Without Protected Attributes: Latent Concept Erasure from Textual Profiles
H-SAL erases latent concepts from text profiles using self-descriptions as implicit debiasing signals and shows competitive performance on a new multi-domain Stack Exchange helpfulness benchmark.
-
Synthetic Sources?: Auditing Generative Search Engine Citations for Evidence of AI-Generated Sources
Audit of ChatGPT, Copilot, Gemini and Perplexity finds ~16% of cited sources are AI-generated across 712 queries on politics, health and environment.
-
Big AI's Regulatory Capture: Mapping Industry Interference and Government Complicity
Big AI captures regulation through 27 mechanisms in five categories, most commonly via discourse influence and law elusion, often justified by narratives that regulation stifles innovation or serves national interest.
-
Beyond Text Prompts: Precise Concept Erasure through Text-Image Collaboration
TICoE achieves more precise and faithful concept erasure in text-to-image models by collaborating text and image data through a convex manifold and hierarchical learning, outperforming prior methods.
-
A Closer Look at the Existing Risks of Generative AI: Mapping the Who, What, and How of Real-World Incidents
Analysis of 499 generative AI incidents shows use-related failures predominate and frequently harm non-users, producing a distinct risk profile from traditional AI.
-
SEAL: Semantic Aware Image Watermarking
SEAL uses semantic embeddings and locality-sensitive hashing to create distortion-free, database-free watermarks for generative images that are conditioned on content for improved forgery resistance.
-
Vibe Coding in Product Teams: Reconfiguring AI-Assisted Workflows, Prototyping, and Collaboration
Interviews reveal a four-stage vibe coding workflow that accelerates prototyping while introducing tensions between quick efficiency and reflective design intention, plus asymmetries in trust and ownership.
-
Trustworthy LLMs: a Survey and Guideline for Evaluating Large Language Models' Alignment
Survey organizes LLM trustworthiness into seven categories and 29 sub-categories, measures eight sub-categories on popular models, and finds that more aligned models generally score higher but with varying effectiveness.
-
The coordination gap in frontier AI safety policies
Frontier AI safety policies have a structural coordination gap caused by diffuse benefits and concentrated costs, which can be addressed by adapting precommitment and shared response protocols from other high-risk domains.
-
Rethinking Text-to-Image as Semantic-Aware Data Augmentation for Indoor Scene Recognition
Stable Diffusion augments limited indoor scene datasets for better recognition models, and DIRE detects the generated images with 100% accuracy using lightweight classifiers.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.