This work provides an explanatory view of how LLMs can apply moral reasoning to both criticize and defend sexist language. We assessed eight large language models, all of which demonstrated the capability to provide explanations grounded in varying moral perspectives for both critiquing and endorsing views that reflect sexist assumptions. With both human and automatic evaluation, we show that all eight models produce comprehensible and contextually relevant text, which is helpful in understanding diverse views on how sexism is perceived. Also, through analysis of moral foundations cited by LLMs in their arguments, we uncover the diverse ideological perspectives in models’ outputs, with some models aligning more with progressive or conservative views on gender roles and sexism.Based on our observations, we caution against the potential misuse of LLMs to justify sexist language. We also highlight that LLMs can serve as tools for understanding the roots of sexist beliefs and designing well-informed interventions. Given this dual capacity, it is crucial to monitor LLMs and design safety mechanisms for their use in applications that involve sensitive societal topics, such as sexism.

Negative public perceptions of people living in poverty can hamper policies and programs that aim to help the poor. One prominent example of social bias and discrimination against people in need is the persistent association of poverty with criminality. The phenomenon has two facets: first, the belief that poor people are more likely to engage in crime (e.g., stealing, mugging, violence) and second, the view that certain behaviors directly resulting from poverty (e.g., living outside, panhandling) warrant criminal punishment. In this paper, we use large language models (LLMs) to identify examples of crime–poverty association (CPA) in English social media texts. We analyze the online discourse on CPA across eight geographically-diverse countries, and find evidence that the CPA rates are higher within the sample obtained from the U.S. and Canada, as compared to the other countries such as South Africa, despite the latter having higher poverty, criminality, and inequality indexes. We further uncover and analyze the most common themes in CPA posts and find more negative and biased attitudes toward people living in poverty in posts from the U.S. and Canada. These results could partially be explained by cultural factors related to the tendency to overestimate the equality of opportunities and social mobility in the U.S. and Canada. These findings have consequences for policy-making and open a new path of research for poverty mitigation with the focus not only on the redistribution of wealth but also on the mitigation of bias and discrimination against people in need.

For decades, psychologists have been studying stereotypes using specially-designed rating scales to capture people’s beliefs and opinions about different social groups. Now, using NLP tools on extensive collections of text, we have the opportunity to study stereotypes “in the wild” and on a large scale. However, are we truly capturing the same information? In this paper we compare measurements along six psychologically-motivated, stereotype-relevant dimensions (Sociability, Morality, Ability, Assertiveness, Beliefs, and Status) for 10 groups, defined by occupation. We compute these measurements on stereotypical English sentences written by crowd-workers, stereotypical sentences generated by ChatGPT, and more general data collected from social media, and contrast the findings with traditional, survey-based results, as well as a spontaneous word-list generation task. We find that while the correlation with the traditional scales varies across dimensions, the free-text data can be used to specify the particular traits associated with each group, and provide context for numerical survey data.

Following on recent advances in large language models (LLMs) and subsequent chat models, a new wave of large vision–language models (LVLMs) has emerged. Such models can incorporate images as input in addition to text, and perform tasks such as visual question answering, image captioning, story generation, etc. Here, we examine potential gender and racial biases in such systems, based on the perceived characteristics of the people in the input images. To accomplish this, we present a new dataset PAIRS (PArallel Images for eveRyday Scenarios). The PAIRS dataset contains sets of AI-generated images of people, such that the images are highly similar in terms of background and visual content, but differ along the dimensions of gender (man, woman) and race (Black, white). By querying the LVLMs with such images, we observe significant differences in the responses according to the perceived gender or race of the person depicted.

Gender stereotypes are pervasive beliefs about individuals based on their gender that play a significant role in shaping societal attitudes, behaviours, and even opportunities. Recognizing the negative implications of gender stereotypes, particularly in online communications, this study investigates eleven strategies to automatically counteract and challenge these views. We present AI-generated gender-based counter-stereotypes to (self-identified) male and female study participants and ask them to assess their offensiveness, plausibility, and potential effectiveness. The strategies of counter-facts and broadening universals (i.e., stating that anyone can have a trait regardless of group membership) emerged as the most robust approaches, while humour, perspective-taking, counter-examples, and empathy for the speaker were perceived as less effective. Also, the differences in ratings were more pronounced for stereotypes about the different targets than between the genders of the raters. Alarmingly, many AI-generated counter-stereotypes were perceived as offensive and/or implausible. Our analysis and the collected dataset offer foundational insight into counter-stereotype generation, guiding future efforts to develop strategies that effectively challenge gender stereotypes in online interactions.

When harmful social stereotypes are expressed on a public platform, they must be addressed in a way that educates and informs both the original poster and other readers, without causing offence or perpetuating new stereotypes. In this paper, we synthesize findings from psychology and computer science to propose a set of potential counter-stereotype strategies. We then automatically generate such counter-stereotypes using ChatGPT, and analyze their correctness and expected effectiveness at reducing stereotypical associations. We identify the strategies of denouncing stereotypes, warning of consequences, and using an empathetic tone as three promising strategies to be further tested.

With the continuous advancement in unsupervised learning methodologies, text generation has become increasingly pervasive. However, the evaluation of the quality of the generated text remains challenging. Human annotations are expensive and often show high levels of disagreement, in particular for certain tasks characterized by inherent subjectivity, such as translation and summarization.Consequently, the demand for automated metrics that can reliably assess the quality of such generative systems and their outputs has grown more pronounced than ever. In 2023, Eval4NLP organized a shared task dedicated to the automatic evaluation of outputs from two specific categories of generative systems: machine translation and summarization. This evaluation was achieved through the utilization of prompts with Large Language Models. Participating in the summarization evaluation track, we propose an approach that involves prompting LLMs to evaluate six different latent dimensions of summarization quality. In contrast to many previous approaches to summarization assessments, which emphasize lexical overlap with reference text, this method surfaces the importance of correct syntax in summarization evaluation. Our method resulted in the second-highest performance in this shared task, demonstrating its effectiveness as a reference-free evaluation.

While many types of hate speech and online toxicity have been the focus of extensive research in NLP, toxic language stigmatizing poor people has been mostly disregarded. Yet, aporophobia, a social bias against the poor, is a common phenomenon online, which can be psychologically damaging as well as hindering poverty reduction policy measures. We demonstrate that aporophobic attitudes are indeed present in social media and argue that the existing NLP datasets and models are inadequate to effectively address this problem. Efforts toward designing specialized resources and novel socio-technical mechanisms for confronting aporophobia are needed.

Classifiers tend to learn a false causal relationship between an over-represented concept and a label, which can result in over-reliance on the concept and compromised classification accuracy. It is imperative to have methods in place that can compare different models and identify over-reliances on specific concepts. We consider three well-known abusive language classifiers trained on large English datasets and focus on the concept of negative emotions, which is an important signal but should not be learned as a sufficient feature for the label of abuse. Motivated by the definition of global sufficiency, we first examine the unwanted dependencies learned by the classifiers by assessing their accuracy on a challenge set across all decision thresholds. Further, recognizing that a challenge set might not always be available, we introduce concept-based explanation metrics to assess the influence of the concept on the labels. These explanations allow us to compare classifiers regarding the degree of false global sufficiency they have learned between a concept and a label.

We present a novel feature attribution method for explaining text classifiers, and analyze it in the context of hate speech detection. Although feature attribution models usually provide a single importance score for each token, we instead provide two complementary and theoretically-grounded scores – necessity and sufficiency – resulting in more informative explanations. We propose a transparent method that calculates these values by generating explicit perturbations of the input text, allowing the importance scores themselves to be explainable. We employ our method to explain the predictions of different hate speech detection models on the same set of curated examples from a test suite, and show that different values of necessity and sufficiency for identity terms correspond to different kinds of false positive errors, exposing sources of classifier bias against marginalized groups.

Previous works on the fairness of toxic language classifiers compare the output of models with different identity terms as input features but do not consider the impact of other important concepts present in the context. Here, besides identity terms, we take into account high-level latent features learned by the classifier and investigate the interaction between these features and identity terms. For a multi-class toxic language classifier, we leverage a concept-based explanation framework to calculate the sensitivity of the model to the concept of sentiment, which has been used before as a salient feature for toxic language detection. Our results show that although for some classes, the classifier has learned the sentiment information as expected, this information is outweighed by the influence of identity terms as input features. This work is a step towards evaluating procedural fairness, where unfair processes lead to unfair outcomes. The produced knowledge can guide debiasing techniques to ensure that important concepts besides identity terms are well-represented in training datasets.

Robustness of machine learning models on ever-changing real-world data is critical, especially for applications affecting human well-being such as content moderation. New kinds of abusive language continually emerge in online discussions in response to current events (e.g., COVID-19), and the deployed abuse detection systems should be updated regularly to remain accurate. In this paper, we show that general abusive language classifiers tend to be fairly reliable in detecting out-of-domain explicitly abusive utterances but fail to detect new types of more subtle, implicit abuse. Next, we propose an interpretability technique, based on the Testing Concept Activation Vector (TCAV) method from computer vision, to quantify the sensitivity of a trained model to the human-defined concepts of explicit and implicit abusive language, and use that to explain the generalizability of the model on new data, in this case, COVID-related anti-Asian hate speech. Extending this technique, we introduce a novel metric, Degree of Explicitness, for a single instance and show that the new metric is beneficial in suggesting out-of-domain unlabeled examples to effectively enrich the training data with informative, implicitly abusive texts.

In an effort to guarantee that machine learning model outputs conform with human moral values, recent work has begun exploring the possibility of explicitly training models to learn the difference between right and wrong. This is typically done in a bottom-up fashion, by exposing the model to different scenarios, annotated with human moral judgements. One question, however, is whether the trained models actually learn any consistent, higher-level ethical principles from these datasets – and if so, what? Here, we probe the Allen AI Delphi model with a set of standardized morality questionnaires, and find that, despite some inconsistencies, Delphi tends to mirror the moral principles associated with the demographic groups involved in the annotation process. We question whether this is desirable and discuss how we might move forward with this knowledge.

Motivations for methods in explainable artificial intelligence (XAI) often include detecting, quantifying and mitigating bias, and contributing to making machine learning models fairer. However, exactly how an XAI method can help in combating biases is often left unspecified. In this paper, we briefly review trends in explainability and fairness in NLP research, identify the current practices in which explainability methods are applied to detect and mitigate bias, and investigate the barriers preventing XAI methods from being used more widely in tackling fairness issues.

Age-related stereotypes are pervasive in our society, and yet have been under-studied in the NLP community. Here, we present a method for extracting age-related stereotypes from Twitter data, generating a corpus of 300,000 over-generalizations about four contemporary generations (baby boomers, generation X, millennials, and generation Z), as well as “old” and “young” people more generally. By employing word-association metrics, semi-supervised topic modelling, and density-based clustering, we uncover many common stereotypes as reported in the media and in the psychological literature, as well as some more novel findings. We also observe trends consistent with the existing literature, namely that definitions of “young” and “old” age appear to be context-dependent, stereotypes for different generations vary across different topics (e.g., work versus family life), and some age-based stereotypes are distinct from generational stereotypes. The method easily extends to other social group labels, and therefore can be used in future work to study stereotypes of different social categories. By better understanding how stereotypes are formed and spread, and by tracking emerging stereotypes, we hope to eventually develop mitigating measures against such biased statements.

Stereotypical language expresses widely-held beliefs about different social categories. Many stereotypes are overtly negative, while others may appear positive on the surface, but still lead to negative consequences. In this work, we present a computational approach to interpreting stereotypes in text through the Stereotype Content Model (SCM), a comprehensive causal theory from social psychology. The SCM proposes that stereotypes can be understood along two primary dimensions: warmth and competence. We present a method for defining warmth and competence axes in semantic embedding space, and show that the four quadrants defined by this subspace accurately represent the warmth and competence concepts, according to annotated lexicons. We then apply our computational SCM model to textual stereotype data and show that it compares favourably with survey-based studies in the psychological literature. Furthermore, we explore various strategies to counter stereotypical beliefs with anti-stereotypes. It is known that countering stereotypes with anti-stereotypical examples is one of the most effective ways to reduce biased thinking, yet the problem of generating anti-stereotypes has not been previously studied. Thus, a better understanding of how to generate realistic and effective anti-stereotypes can contribute to addressing pressing societal concerns of stereotyping, prejudice, and discrimination.

When comparing entities extracted by a medical entity recognition system with gold standard annotations over a test set, two types of mismatches might occur, label mismatch or span mismatch. Here we focus on span mismatch and show that its severity can vary from a serious error to a fully acceptable entity extraction due to the subjectivity of span annotations. For a domain-specific BERT-based NER system, we showed that 25% of the errors have the same labels and overlapping span with gold standard entities. We collected expert judgement which shows more than 90% of these mismatches are accepted or partially accepted by the user. Using the training set of the NER system, we built a fast and lightweight entity classifier to approximate the user experience of such mismatches through accepting or rejecting them. The decisions made by this classifier are used to calculate a learning-based F-score which is shown to be a better approximation of a forgiving user’s experience than the relaxed F-score. We demonstrated the results of applying the proposed evaluation metric for a variety of deep learning medical entity recognition models trained with two datasets.

There is growing evidence that changes in speech and language may be early markers of dementia, but much of the previous NLP work in this area has been limited by the size of the available datasets. Here, we compare several methods of domain adaptation to augment a small French dataset of picture descriptions (n = 57) with a much larger English dataset (n = 550), for the task of automatically distinguishing participants with dementia from controls. The first challenge is to identify a set of features that transfer across languages; in addition to previously used features based on information units, we introduce a new set of features to model the order in which information units are produced by dementia patients and controls. These concept-based language model features improve classification performance in both English and French separately, and the best result (AUC = 0.89) is achieved using the multilingual training set with a combination of information and language model features.

Entity recognition is a critical first step to a number of clinical NLP applications, such as entity linking and relation extraction. We present the first attempt to apply state-of-the-art entity recognition approaches on a newly released dataset, MedMentions. This dataset contains over 4000 biomedical abstracts, annotated for UMLS semantic types. In comparison to existing datasets, MedMentions contains a far greater number of entity types, and thus represents a more challenging but realistic scenario in a real-world setting. We explore a number of relevant dimensions, including the use of contextual versus non-contextual word embeddings, general versus domain-specific unsupervised pre-training, and different deep learning architectures. We contrast our results against the well-known i2b2 2010 entity recognition dataset, and propose a new method to combine general and domain-specific information. While producing a state-of-the-art result for the i2b2 2010 task (F1 = 0.90), our results on MedMentions are significantly lower (F1 = 0.63), suggesting there is still plenty of opportunity for improvement on this new data.

In 2014, a chatty but immobile robot called hitchBOT set out to hitchhike across Canada. It similarly made its way across Germany and the Netherlands, and had begun a trip across the USA when it was destroyed by vandals. In this work, we analyze the emotions and sentiments associated with words in tweets posted before and after hitchBOT’s destruction to answer two questions: Were there any differences in the emotions expressed across the different countries visited by hitchBOT? And how did the public react to the demise of hitchBOT? Our analyses indicate that while there were few cross-cultural differences in sentiment towards hitchBOT, there was a significant negative emotional reaction to its destruction, suggesting that people had formed an emotional connection with hitchBOT and perceived its destruction as morally wrong. We discuss potential implications of anthropomorphism and emotional attachment to robots from the perspective of robot ethics.

Increased access to large datasets has driven progress in NLP. However, most computational studies of clinically-validated, patient-generated speech and language involve very few datapoints, as such data are difficult (and expensive) to collect. In this position paper, we argue that we must find ways to promote data sharing across research groups, in order to build datasets of a more appropriate size for NLP and machine learning analysis. We review the benefits and challenges of sharing clinical language data, and suggest several concrete actions by both clinical and NLP researchers to encourage multi-site and multi-disciplinary data sharing. We also propose the creation of a collaborative data sharing platform, to allow NLP researchers to take a more active responsibility for data transcription, annotation, and curation.

We present a machine learning analysis of eye-tracking data for the detection of mild cognitive impairment, a decline in cognitive abilities that is associated with an increased risk of developing dementia. We compare two experimental configurations (reading aloud versus reading silently), as well as two methods of combining information from the two trials (concatenation and merging). Additionally, we annotate the words being read with information about their frequency and syntactic category, and use these annotations to generate new features. Ultimately, we are able to distinguish between participants with and without cognitive impairment with up to 86% accuracy.