Misogynoir: challenges in detecting intersectional hate

Kwarteng, Joseph; Perfumi, Serena Coppolino; Farrell, Tracie; Third, Allan; Fernández, Miriam

doi:10.1007/s13278-022-00993-7

Cited by 10 publications

(4 citation statements)

References 47 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Building on algorithms that extract personality profiles from online activity ( 23 , 24 ), there are now AI tools that attempt to automatically score individual social media users on traits such as sexism or aggression ( 25–27 ), which can then be used to form a moral impression of these users. AI moral scoring also fuels large-scale social engineering projects such as the Chinese social credit system, in which a wide range of behaviors are aggregated into a single score for every citizen, with social and legal penalties for citizens who drop beneath a certain score.…”

Section: Introductionmentioning

confidence: 99%

Humans feel too special for machines to score their morals

Purcell¹,

Bonnefon

2023

PNAS Nexus

View full text Add to dashboard Cite

Artificial Intelligence (AI) can be harnessed to create sophisticated social and moral scoring systems—enabling people and organizations to form judgements of others at scale. However, it also poses significant ethical challenges and is, subsequently, the subject of wide debate. As these technologies are developed and governing bodies face regulatory decisions, it is crucial that we understand the attraction or resistance that people have for AI moral scoring. Across four experiments, we show that the acceptability of moral scoring by AI is related to expectations about the quality of those scores, but that expectations about quality are compromised by people's tendency to see themselves as morally peculiar. We demonstrate that people overestimate the peculiarity of their moral profile, believe that AI will neglect this peculiarity, and resist for this reason the introduction of moral scoring by AI.

show abstract

Section: Introductionmentioning

confidence: 99%

Humans feel too special for machines to score their morals

Purcell¹,

Bonnefon

2023

PNAS Nexus

View full text Add to dashboard Cite

show abstract

“…The analysis of sentiment and detection of biases enabled the development of tools to counteract biased reporting, enhancing the objectivity of news dissemination [30,31,32]. LLMs' ability to process and interpret natural language allowed for the identification of subtle emotional undertones that could influence public perception [33,34,35]. The application of LLMs in sentiment analysis facilitated the assessment of the overall tone of news articles, providing insights into the potential biases present in the content [36,37].…”

Section: Sentiment Analysis and Bias Detectionmentioning

confidence: 99%

Fake News Detection with Large Language Models on the LIAR Dataset

Boissonneault,

Hensen

2024

Preprint

View full text Add to dashboard Cite

The widespread dissemination of fake news poses a significant threat to the integrity of information. Detecting fake news with high accuracy is crucial for maintaining the integrity of information in the digital age. The evaluation of ChatGPT and Google Gemini models for this task has revealed their substantial capabilities in discerning the veracity of statements, highlighting their potential to mitigate the spread of misinformation. Using the LIAR benchmark dataset, the study demonstrated high performance metrics across accuracy, precision, recall, F1 score, and AUC-ROC, emphasizing the effectiveness of these models in real-world applications. The comparative analysis and error examination provided insights into the strengths and limitations of each model, offering valuable guidance for future enhancements. Practical implications include the integration of these models into fact-checking systems to improve content verification processes, supporting media organizations and social platforms in their efforts to combat misinformation. The findings prove the importance of ongoing research and development to refine and optimize LLMs, ensuring their continued relevance and efficacy in addressing the challenges posed by fake news.

show abstract

“…Content moderation systems generally focus on defining policies to protect any identity group or individual targeted [13]. Nevertheless, the specific sociolinguistic aspects of harmful expressions [14,15] make this phenomenon different for each target. A system focused on recognising hate directed to a specific group would not generalise to a different identity [2].…”

Section: Related Workmentioning

confidence: 99%

Knowledge-Grounded Target Group Language Recognition in Hate Speech

Reyero Lobo,

Daga,

Alani

et al. 2023

Knowledge Graphs: Semantics, Machine Learning, and Languages

View full text Add to dashboard Cite

Hate speech comes in different forms depending on the communities targeted, often based on factors like gender, sexuality, race, or religion. Detecting it online is challenging because existing systems are not accounting for the diversity of hate based on the identity of the target and may be biased towards certain groups, leading to inaccurate results. Current language models perform well in identifying target communities, but only provide a probability that a hate speech text contains references to a particular group. This lack of transparency is problematic because these models learn biases from data annotated by individuals who may not be familiar with the target group. To improve hate speech detection, particularly target group identification, we propose a new hybrid approach that incorporates explicit knowledge about the language used by specific identity groups. We leverage a Knowledge Graph (KG) and adapt it, considering an appropriate level of abstraction, to recognise hate speech-language related to gender and sexual orientation. A thorough quantitative and qualitative evaluation demonstrates that our approach is as effective as state-of-the-art language models while adjusting better to domain and data changes. By grounding the task in explicit knowledge, we can better contextualise the results generated by our proposed approach with the language of the groups most frequently impacted by these technologies. Semantic enrichment helps us examine model outcomes and the training data used for hate speech detection systems, and handle ambiguous cases in human annotations more effectively. Overall, infusing semantic knowledge in hate speech detection is crucial for enhancing understanding of model behaviors and addressing biases derived from training data.

show abstract

Misogynoir: challenges in detecting intersectional hate

Cited by 10 publications

References 47 publications

Humans feel too special for machines to score their morals

Humans feel too special for machines to score their morals

Fake News Detection with Large Language Models on the LIAR Dataset

Knowledge-Grounded Target Group Language Recognition in Hate Speech

Contact Info

Product

Resources

About