A Framework for Automated Measurement of Responsible AI Harms in Generative AI Applications

We present a framework for the automated measurement of responsible AI (RAI) metrics for large language models (LLMs) and associated products and services. Our framework for automatically measuring harms from LLMs builds on existing technical and sociotechnical expertise and leverages the capabilities of state-of-the-art LLMs, such as GPT-4. We use this framework to run through several case studies investigating how different LLMs may violate a range of RAI-related principles. The framework may be employed alongside domain-specific sociotechnical expertise to create measurements for new harm areas in the future. By implementing this framework, we aim to enable more advanced harm measurement efforts and further the responsible use of LLMs.

A Keyword Based Approach to Understanding the Overpenalization of Marginalized Groups by English Marginal Abuse Models on Twitter

Harmful content detection models tend to have higher false positive rates for content from marginalized groups. In the context of marginal abuse modeling on Twitter, such disproportionate penalization poses the risk of reduced visibility, where marginalized communities lose the opportunity to voice their opinion on the platform. Current approaches to algorithmic harm mitigation, and bias detection for NLP models are often very ad hoc and subject to human bias. We make two main contributions in this paper. First, we design a novel methodology, which provides a principled approach to detecting and measuring the severity of potential harms associated with a text-based model. Second, we apply our methodology to audit Twitter's English marginal abuse model, which is used for removing amplification eligibility of marginally abusive content. Without utilizing demographic labels or dialect classifiers, we are still able to detect and measure the severity of issues related to the over-penalization of the speech of marginalized communities, such as the use of reclaimed speech, counterspeech, and identity related terms. In order to mitigate the associated harms, we experiment with adding additional true negative examples and find that doing so provides improvements to our fairness metrics without large degradations in model performance.

What do Bias Measures Measure?

Natural Language Processing (NLP) models propagate social biases about protected attributes such as gender, race, and nationality. To create interventions and mitigate these biases and associated harms, it is vital to be able to detect and measure such biases. While many existing works propose bias evaluation methodologies for different tasks, there remains a need to cohesively understand what biases and normative harms each of these measures captures and how different measures compare. To address this gap, this work presents a comprehensive survey of existing bias measures in NLP as a function of the associated NLP tasks, metrics, datasets, and social biases and corresponding harms. This survey also organizes metrics into different categories to present advantages and disadvantages. Finally, we propose a documentation standard for bias measures to aid their development, categorization, and appropriate usage.

Societal Biases in Language Generation: Progress and Challenges

Technology for language generation has advanced rapidly, spurred by advancements in pre-training large models on massive amounts of data and the need for intelligent agents to communicate in a natural manner. While techniques can effectively generate fluent text, they can also produce undesirable societal biases that can have a disproportionately negative impact on marginalized populations. Language generation presents unique challenges for biases in terms of direct user interaction and the structure of decoding techniques. To better understand these challenges, we present a survey on societal biases in language generation, focusing on how data and techniques contribute to biases and progress towards reducing biases. Motivated by a lack of studies on biases from decoding techniques, we also conduct experiments to quantify the effects of these techniques. By further discussing general trends and open challenges, we call to attention promising directions for research and the importance of fairness and inclusivity considerations for language generation applications.

Revealing Persona Biases in Dialogue Systems

Dialogue systems in the form of chatbots and personal assistants are being increasingly integrated into people's lives. These dialogue systems often have the ability to adopt an anthropomorphic persona, mimicking a societal demographic to appear more approachable and trustworthy to users. However, the adoption of a persona can result in the adoption of biases. We define persona biases as harmful differences in text (e.g., varying levels of offensiveness or affirmations of biased statements) generated from adopting different demographic personas. In this paper, we present the first large-scale study on persona biases in dialogue systems and conduct analyses on personas of different social classes, sexual orientations, races, and genders. Furthermore, we introduce an open-source framework, UnitPersonaBias, a tool to explore and aggregate subtle persona biases in dialogue systems. In our studies of the Blender and DialoGPT dialogue systems, we show that the choice of personas can affect the degree of harms in generated responses. Additionally, adopting personas of more diverse, historically marginalized demographics appears to decrease harmful responses the most.

Investigating Societal Biases in a Poetry Composition System

There is a growing collection of work analyzing and mitigating societal biases in language understanding, generation, and retrieval tasks, though examining biases in creative tasks remains underexplored. Creative language applications are meant for direct interaction with users, so it is important to quantify and mitigate societal biases in these applications. We introduce a novel study on a pipeline to mitigate societal biases when retrieving next verse suggestions in a poetry composition system. Our results suggest that data augmentation through sentiment style transfer has potential for mitigating societal biases.

"Nice Try, Kiddo": Ad Hominems in Dialogue Systems

Ad hominem attacks are those that attack some feature of a person's character instead of the position the person is maintaining. As a form of toxic and abusive language, ad hominems contain harmful language that could further amplify the skew of power inequality for marginalized populations. Since dialogue systems are designed to respond directly to user input, it is important to study ad hominems in these system responses. In this work, we propose categories of ad hominems that allow us to analyze human and dialogue system responses to Twitter posts. We specifically compare responses to Twitter posts about marginalized communities (#BlackLivesMatter, #MeToo) and other topics (#Vegan, #WFH). Furthermore, we propose a constrained decoding technique that uses salient $n$-gram similarity to apply soft constraints to top-$k$ sampling and can decrease the amount of ad hominems generated by dialogue systems. Our results indicate that 1) responses composed by both humans and DialoGPT contain more ad hominems for discussions around marginalized communities versus other topics, 2) different amounts of ad hominems in the training data can influence the likelihood of the model generating ad hominems, and 3) we can thus carefully choose training data and use constrained decoding techniques to decrease the amount of ad hominems generated by dialogue systems.

Towards Controllable Biases in Language Generation

We present a general approach towards controllable societal biases in natural language generation (NLG). Building upon the idea of adversarial triggers, we develop a method to induce or avoid biases in generated text containing mentions of specified demographic groups. We then analyze two scenarios: 1) inducing biases for one demographic and avoiding biases for another, and 2) mitigating biases between demographic pairs (e.g., man and woman). The former scenario gives us a tool for detecting the types of biases present in the model, and the latter is useful for mitigating biases in downstream applications (e.g., dialogue generation). Specifically, our approach facilitates more explainable biases by allowing us to 1) use the relative effectiveness of inducing biases for different demographics as a new dimension for bias evaluation, and 2) discover topics that correspond to demographic inequalities in generated text. Furthermore, our mitigation experiments exemplify our technique's effectiveness at equalizing the amount of biases across demographics while simultaneously generating less negatively biased text overall.

The Woman Worked as a Babysitter: On Biases in Language Generation

We present a systematic study of biases in natural language generation (NLG) by analyzing text generated from prompts that contain mentions of different demographic groups. In this work, we introduce the notion of the regard towards a demographic, use the varying levels of regard towards different demographics as a defining metric for bias in NLG, and analyze the extent to which sentiment scores are a relevant proxy metric for regard. To this end, we collect strategically-generated text from language models and manually annotate the text with both sentiment and regard scores. Additionally, we build an automatic regard classifier through transfer learning, so that we can analyze biases in unseen text. Together, these methods reveal the extent of the biased nature of language model generations. Our analysis provides a study of biases in NLG, bias metrics and correlated human judgments, and empirical evidence on the usefulness of our annotated dataset.