Evaluating the Social Impact of Generative AI Systems in Systems and Society

By Irene Solaiman, Zeerak Talat, William Agnew, L. Ahmad, Dylan Baker, Su Lin Blodgett, Canyu Chen, Hal Daum'e, Jesse Dodge, Ellie Evans and 13 more. 2023

177 citations recorded by the source corpus.

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.

Subfields: Artificial Intelligence, Safety Research, Safety, Robustness & Interpretability

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Safety Research

In Safety, Robustness & Interpretability

Citations in the corpus

Llama 2: Open Foundation and Fine-Tuned Chat Models (2023, background): “Like all LLMs, Llama 2 is a new technology that carries potential risks with use (Bender et al., 2021b; Weidinger et al., 2021; Solaiman et al., 2023).”
GPT-4o System Card (2024, background): “Researchers at OpenAI and elsewhere have discussed a range of possible impacts, from societal harms (including representational harms [18, 12, 23, 24]; disinformation, misinformation, and influence operations [18, 25, 23], environmental harms [12, 23], attachment [26], misuse [27, 23], and loss of…”
Knowledge Conflicts for LLMs: A Survey (2024, background): “Several studies highlight the challenges faced by LMs in identifying contradictions.”
Managing extreme AI risks amid rapid progress (2023, background): “Especially in the hands of a few powerful actors, AI could cement or exacerbate global inequities, or facilitate automated warfare, customized mass manipulation, and pervasive surveillance 18–23 .”