HateDay: Insights from a Global Hate Speech Dataset Representative of a Day on Twitter

To tackle the global challenge of online hate speech, a large body of research has developed detection models to flag hate speech in the sea of online content. Yet, due to systematic biases in evaluation datasets, detection performance in real-world settings remains unclear, let alone across geographies. To address this issue, we introduce HateDay, the first global hate speech dataset representative of social media settings, randomly sampled from all tweets posted on September 21, 2022 for eight languages and four English-speaking countries. Using HateDay, we show how the prevalence and composition of hate speech varies across languages and countries. We also find that evaluation on academic hate speech datasets overestimates real-world detection performance, which we find is very low, especially for non-European languages. We identify several factors explaining poor performance, including models' inability to distinguish between hate and offensive speech, and the misalignment between academic target focus and real-world target prevalence. We finally argue that such low performance renders hate speech moderation with public detection models unfeasible, even in a human-in-the-loop setting which we find is prohibitively costly. Overall, we emphasize the need to evaluate future detection models from academia and platforms in real-world settings to address this global challenge.

LINGOLY: A Benchmark of Olympiad-Level Linguistic Reasoning Puzzles in Low-Resource and Extinct Languages

In this paper, we present the LingOly benchmark, a novel benchmark for advanced reasoning abilities in large language models. Using challenging Linguistic Olympiad puzzles, we evaluate (i) capabilities for in-context identification and generalisation of linguistic patterns in very low-resource or extinct languages, and (ii) abilities to follow complex task instructions. The LingOly benchmark covers more than 90 mostly low-resource languages, minimising issues of data contamination, and contains 1,133 problems across 6 formats and 5 levels of human difficulty. We assess performance with both direct accuracy and comparison to a no-context baseline to penalise memorisation. Scores from 11 state-of-the-art LLMs demonstrate the benchmark to be challenging, and models perform poorly on the higher difficulty problems. On harder problems, even the top model only achieved 38.7% accuracy, 24.7% improvement over the no-context baseline. Large closed models typically outperform open models, and in general, the higher resource the language, the better the scores. These results indicate, in absence of memorisation, true multi-step out-of-domain reasoning remains a challenge for current language models.

SynDy: Synthetic Dynamic Dataset Generation Framework for Misinformation Tasks

Diaspora communities are disproportionately impacted by off-the-radar misinformation and often neglected by mainstream fact-checking efforts, creating a critical need to scale-up efforts of nascent fact-checking initiatives. In this paper we present SynDy, a framework for Synthetic Dynamic Dataset Generation to leverage the capabilities of the largest frontier Large Language Models (LLMs) to train local, specialized language models. To the best of our knowledge, SynDy is the first paper utilizing LLMs to create fine-grained synthetic labels for tasks of direct relevance to misinformation mitigation, namely Claim Matching, Topical Clustering, and Claim Relationship Classification. SynDy utilizes LLMs and social media queries to automatically generate distantly-supervised, topically-focused datasets with synthetic labels on these three tasks, providing essential tools to scale up human-led fact-checking at a fraction of the cost of human-annotated data. Training on SynDy's generated labels shows improvement over a standard baseline and is not significantly worse compared to training on human labels (which may be infeasible to acquire). SynDy is being integrated into Meedan's chatbot tiplines that are used by over 50 organizations, serve over 230K users annually, and automatically distribute human-written fact-checks via messaging apps such as WhatsApp. SynDy will also be integrated into our deployed Co-Insights toolkit, enabling low-resource organizations to launch tiplines for their communities. Finally, we envision SynDy enabling additional fact-checking tools such as matching new misinformation claims to high-quality explainers on common misinformation topics.

Global News Synchrony and Diversity During the Start of the COVID-19 Pandemic

News coverage profoundly affects how countries and individuals behave in international relations. Yet, we have little empirical evidence of how news coverage varies across countries. To enable studies of global news coverage, we develop an efficient computational methodology that comprises three components: (i) a transformer model to estimate multilingual news similarity; (ii) a global event identification system that clusters news based on a similarity network of news articles; and (iii) measures of news synchrony across countries and news diversity within a country, based on country-specific distributions of news coverage of the global events. Each component achieves state-of-the art performance, scaling seamlessly to massive datasets of millions of news articles. We apply the methodology to 60 million news articles published globally between January 1 and June 30, 2020, across 124 countries and 10 languages, detecting 4357 news events. We identify the factors explaining diversity and synchrony of news coverage across countries. Our study reveals that news media tend to cover a more diverse set of events in countries with larger Internet penetration, more official languages, larger religious diversity, higher economic inequality, and larger populations. Coverage of news events is more synchronized between countries that not only actively participate in commercial and political relations -- such as, pairs of countries with high bilateral trade volume, and countries that belong to the NATO military alliance or BRICS group of major emerging economies -- but also countries that share certain traits: an official language, high GDP, and high democracy indices.

From Languages to Geographies: Towards Evaluating Cultural Bias in Hate Speech Datasets

Perceptions of hate can vary greatly across cultural contexts. Hate speech (HS) datasets, however, have traditionally been developed by language. This hides potential cultural biases, as one language may be spoken in different countries home to different cultures. In this work, we evaluate cultural bias in HS datasets by leveraging two interrelated cultural proxies: language and geography. We conduct a systematic survey of HS datasets in eight languages and confirm past findings on their English-language bias, but also show that this bias has been steadily decreasing in the past few years. For three geographically-widespread languages -- English, Arabic and Spanish -- we then leverage geographical metadata from tweets to approximate geo-cultural contexts by pairing language and country information. We find that HS datasets for these languages exhibit a strong geo-cultural bias, largely overrepresenting a handful of countries (e.g., US and UK for English) relative to their prominence in both the broader social media population and the general population speaking these languages. Based on these findings, we formulate recommendations for the creation of future HS datasets.

The PRISM Alignment Project: What Participatory, Representative and Individualised Human Feedback Reveals About the Subjective and Multicultural Alignment of Large Language Models

Human feedback plays a central role in the alignment of Large Language Models (LLMs). However, open questions remain about the methods (how), domains (where), people (who) and objectives (to what end) of human feedback collection. To navigate these questions, we introduce PRISM, a new dataset which maps the sociodemographics and stated preferences of 1,500 diverse participants from 75 countries, to their contextual preferences and fine-grained feedback in 8,011 live conversations with 21 LLMs. PRISM contributes (i) wide geographic and demographic participation in human feedback data; (ii) two census-representative samples for understanding collective welfare (UK and US); and (iii) individualised feedback where every rating is linked to a detailed participant profile, thus permitting exploration of personalisation and attribution of sample artefacts. We focus on collecting conversations that centre subjective and multicultural perspectives on value-laden and controversial topics, where we expect the most interpersonal and cross-cultural disagreement. We demonstrate the usefulness of PRISM via three case studies of dialogue diversity, preference diversity, and welfare outcomes, showing that it matters which humans set alignment norms. As well as offering a rich community resource, we advocate for broader participation in AI development and a more inclusive approach to technology design.

Introducing v0.5 of the AI Safety Benchmark from MLCommons

This paper introduces v0.5 of the AI Safety Benchmark, which has been created by the MLCommons AI Safety Working Group. The AI Safety Benchmark has been designed to assess the safety risks of AI systems that use chat-tuned language models. We introduce a principled approach to specifying and constructing the benchmark, which for v0.5 covers only a single use case (an adult chatting to a general-purpose assistant in English), and a limited set of personas (i.e., typical users, malicious users, and vulnerable users). We created a new taxonomy of 13 hazard categories, of which 7 have tests in the v0.5 benchmark. We plan to release version 1.0 of the AI Safety Benchmark by the end of 2024. The v1.0 benchmark will provide meaningful insights into the safety of AI systems. However, the v0.5 benchmark should not be used to assess the safety of AI systems. We have sought to fully document the limitations, flaws, and challenges of v0.5. This release of v0.5 of the AI Safety Benchmark includes (1) a principled approach to specifying and constructing the benchmark, which comprises use cases, types of systems under test (SUTs), language and context, personas, tests, and test items; (2) a taxonomy of 13 hazard categories with definitions and subcategories; (3) tests for seven of the hazard categories, each comprising a unique set of test items, i.e., prompts. There are 43,090 test items in total, which we created with templates; (4) a grading system for AI systems against the benchmark; (5) an openly available platform, and downloadable tool, called ModelBench that can be used to evaluate the safety of AI systems on the benchmark; (6) an example evaluation report which benchmarks the performance of over a dozen openly available chat-tuned language models; (7) a test specification for the benchmark.

SimpleSafetyTests: a Test Suite for Identifying Critical Safety Risks in Large Language Models

The past year has seen rapid acceleration in the development of large language models (LLMs). For many tasks, there is now a wide range of open-source and open-access LLMs that are viable alternatives to proprietary models like ChatGPT. Without proper steering and safeguards, however, LLMs will readily follow malicious instructions, provide unsafe advice, and generate toxic content. This is a critical safety risk for businesses and developers. We introduce SimpleSafetyTests as a new test suite for rapidly and systematically identifying such critical safety risks. The test suite comprises 100 test prompts across five harm areas that LLMs, for the vast majority of applications, should refuse to comply with. We test 11 popular open LLMs and find critical safety weaknesses in several of them. While some LLMs do not give a single unsafe response, most models we test respond unsafely on more than 20% of cases, with over 50% unsafe responses in the extreme. Prepending a safety-emphasising system prompt substantially reduces the occurrence of unsafe responses, but does not completely stop them from happening. We recommend that developers use such system prompts as a first line of defence against critical safety risks.

Lost in Translation -- Multilingual Misinformation and its Evolution

Misinformation and disinformation are growing threats in the digital age, spreading rapidly across languages and borders. This paper investigates the prevalence and dynamics of multilingual misinformation through an analysis of over 250,000 unique fact-checks spanning 95 languages. First, we find that while the majority of misinformation claims are only fact-checked once, 11.7%, corresponding to more than 21,000 claims, are checked multiple times. Using fact-checks as a proxy for the spread of misinformation, we find 33% of repeated claims cross linguistic boundaries, suggesting that some misinformation permeates language barriers. However, spreading patterns exhibit strong homophily, with misinformation more likely to spread within the same language. To study the evolution of claims over time and mutations across languages, we represent fact-checks with multilingual sentence embeddings and cluster semantically similar claims. We analyze the connected components and shortest paths connecting different versions of a claim finding that claims gradually drift over time and undergo greater alteration when traversing languages. Overall, this novel investigation of multilingual misinformation provides key insights. It quantifies redundant fact-checking efforts, establishes that some claims diffuse across languages, measures linguistic homophily, and models the temporal and cross-lingual evolution of claims. The findings advocate for expanded information sharing between fact-checkers globally while underscoring the importance of localized verification.

The Past, Present and Better Future of Feedback Learning in Large Language Models for Subjective Human Preferences and Values

Human feedback is increasingly used to steer the behaviours of Large Language Models (LLMs). However, it is unclear how to collect and incorporate feedback in a way that is efficient, effective and unbiased, especially for highly subjective human preferences and values. In this paper, we survey existing approaches for learning from human feedback, drawing on 95 papers primarily from the ACL and arXiv repositories.First, we summarise the past, pre-LLM trends for integrating human feedback into language models. Second, we give an overview of present techniques and practices, as well as the motivations for using feedback; conceptual frameworks for defining values and preferences; and how feedback is collected and from whom. Finally, we encourage a better future of feedback learning in LLMs by raising five unresolved conceptual and practical challenges.