AI-infused search engines from Google, Microsoft, and Perplexity have all been surfacing deeply racist and widely debunked research promoting race science and the idea that whites are genetically superior to nonwhites. Patrik Hermansson, a researcher with UK-based anti-racism group Hope Not Hate, was in the middle of a months-long investigation into the resurgent race science movement when he needed to find out some more information about a debunked dataset that claims IQ scores can be used to prove the superiority of the white race. Hermansson was investigating the Human Diversity Foundation, a race science company funded by Andrew Conru, the US tech billionaire who founded Adult Friend Finder. The group, founded in 2022, was the successor to the Pioneer Fund, a group founded by US Nazi sympathizers in 1937 with the aim of promoting "race betterment" and "race realism." Hermansson logged onto Google and began looking up results for the IQs of different nations.

The mother of a teenage boy in the United States who took his own life is suing the maker of an artificial intelligence-powered chatbot that she claims encouraged her son's death. In a lawsuit filed in Florida, Megan Garcia, whose 14-year-old son Sewell Setzer died by suicide in February, accuses Character.AI of complicity in her son's death after he developed a virtual relationship with a chatbot based on the identity of "Game of Thrones" character Daenerys Targaryen. Character.AI's chatbot targeted the teen with "hypersexualized" and "frighteningly realistic experiences" and repeatedly raised the topic of suicide after he had expressed suicidal thoughts, according to the lawsuit filed in Orlando on Tuesday. The lawsuit alleges the chatbot posed as a licensed therapist, encouraging the teen's suicidal ideation and engaging in sexualised conversations that would count as abuse if initiated by a human adult. In his last conversation with the AI before his death, Setzer said he loved the chatbot and would "come home to you", according to the lawsuit.

The world is evolving and so is the vocabulary used to discuss topics in speech. Analysing political speech data from more than 30 years requires the use of flexible topic models to uncover the latent topics and their change in prevalence over time as well as the change in the vocabulary of the topics. We propose the temporal Poisson factorisation (TPF) model as an extension to the Poisson factorisation model to model sparse count data matrices obtained based on the bag-of-words assumption from text documents with time stamps. We discuss and empirically compare different model specifications for the time-varying latent variables consisting either of a flexible auto-regressive structure of order one or a random walk. Estimation is based on variational inference where we consider a combination of coordinate ascent updates with automatic differentiation using batching of documents. Suitable variational families are proposed to ease inference. We compare results obtained using independent univariate variational distributions for the time-varying latent variables to those obtained with a multivariate variant. We discuss in detail the results of the TPF model when analysing speeches from 18 sessions in the U.S. Senate (1981-2016).

Large Language Models (LLMs) suffer from huge number of parameters, which restricts their deployment on edge devices. Weight sharing is one promising solution that encourages weight reuse, effectively reducing memory usage with less performance drop. However, current weight sharing techniques primarily focus on small-scale models like BERT and employ coarse-grained sharing rules, e.g., layer-wise. This becomes limiting given the prevalence of LLMs and sharing an entire layer or block obviously diminishes the flexibility of weight sharing. In this paper, we present a perspective on head-wise shareable attention for large language models. We further propose two memory-efficient methods that share parameters across attention heads, with a specific focus on LLMs. Both of them use the same dynamic strategy to select the shared weight matrices. The first method directly reuses the pre-trained weights without retraining, denoted as $\textbf{DirectShare}$. The second method first post-trains with constraint on weight matrix similarity and then shares, denoted as $\textbf{PostShare}$. Experimental results reveal our head-wise shared models still maintain satisfactory capabilities, demonstrating the feasibility of fine-grained weight sharing applied to LLMs.

Large Language Models (LLMs) have gained significant attention in the medical domain for their human-level capabilities, leading to increased efforts to explore their potential in various healthcare applications. However, despite such a promising future, there are multiple challenges and obstacles that remain for their real-world uses in practical settings. This work discusses key challenges for LLMs in medical applications from four unique aspects: operational vulnerabilities, ethical and social considerations, performance and assessment difficulties, and legal and regulatory compliance. Addressing these challenges is crucial for leveraging LLMs to their full potential and ensuring their responsible integration into healthcare.

With the rapid development of Large language models (LLMs), understanding the capabilities of LLMs in identifying unsafe content has become increasingly important. While previous works have introduced several benchmarks to evaluate the safety risk of LLMs, the community still has a limited understanding of current LLMs' capability to recognize illegal and unsafe content in Chinese contexts. In this work, we present a Chinese safety benchmark (ChineseSafe) to facilitate research on the content safety of large language models. To align with the regulations for Chinese Internet content moderation, our ChineseSafe contains 205,034 examples across 4 classes and 10 sub-classes of safety issues. For Chinese contexts, we add several special types of illegal content: political sensitivity, pornography, and variant/homophonic words. Moreover, we employ two methods to evaluate the legal risks of popular LLMs, including open-sourced models and APIs. The results reveal that many LLMs exhibit vulnerability to certain types of safety issues, leading to legal risks in China. Our work provides a guideline for developers and researchers to facilitate the safety of LLMs.

The rapid advancement of large language models (LLMs) has revolutionized the field of natural language processing, enabling the development of increasingly sophisticated AI systems capable of generating human-like text, engaging in dialogue, and performing complex language tasks [5]. As these models grow in size and capability, the challenge of accurately evaluating their performance and aligning their outputs with human preferences has become increasingly critical [3, 15, 49]. Traditional evaluation methods, such as human assessments and automated metrics, often struggle to capture the nuances and complexities of LLM outputs, leading to a gap between model performance and user expectations [7, 17, 24]. Human evaluations are time-consuming, expensive, and prone to inconsistency and bias [12, 27], while automated metrics frequently fail to align with human judgments, particularly in open-ended generation tasks [29, 13, 22]. To address these challenges, we propose a novel framework for evaluating LLM outputs using LLMs themselves as interacting agents in a courtroom-inspired, multi-agent system. Our approach draws inspiration from various fields, including decision theory, economics, psychology, legal theory, and voting theory, to develop a more dynamic, contextual, and comprehensive assessment process.

Natural Language Processing is revolutionizing the way legal professionals and laypersons operate in the legal field. The considerable potential for Natural Language Processing in the legal sector, especially in developing computational tools for various legal processes, has captured the interest of researchers for years. This survey follows the Preferred Reporting Items for Systematic Reviews and Meta-Analyses framework, reviewing 148 studies, with a final selection of 127 after manual filtering. It explores foundational concepts related to Natural Language Processing in the legal domain, illustrating the unique aspects and challenges of processing legal texts, such as extensive document length, complex language, and limited open legal datasets. We provide an overview of Natural Language Processing tasks specific to legal text, such as Legal Document Summarization, legal Named Entity Recognition, Legal Question Answering, Legal Text Classification, and Legal Judgment Prediction. In the section on legal Language Models, we analyze both developed Language Models and approaches for adapting general Language Models to the legal domain. Additionally, we identify 15 Open Research Challenges, including bias in Artificial Intelligence applications, the need for more robust and interpretable models, and improving explainability to handle the complexities of legal language and reasoning.

The emergence of Segment Anything (SAM) sparked research interest in the field of interactive segmentation, especially in the context of image editing tasks and speeding up data annotation. Unlike common semantic segmentation, interactive segmentation methods allow users to directly influence their output through prompts (e.g. clicks). However, click patterns in real-world interactive segmentation scenarios remain largely unexplored. Most methods rely on the assumption that users would click in the center of the largest erroneous area. Nevertheless, recent studies show that this is not always the case. Thus, methods may have poor performance in real-world deployment despite high metrics in a baseline benchmark. To accurately simulate real-user clicks, we conducted a large crowdsourcing study of click patterns in an interactive segmentation scenario and collected 475K real-user clicks. Drawing on ideas from saliency tasks, we develop a clickability model that enables sampling clicks, which closely resemble actual user inputs. Using our model and dataset, we propose RClicks benchmark for a comprehensive comparison of existing interactive segmentation methods on realistic clicks. Specifically, we evaluate not only the average quality of methods, but also the robustness w.r.t. click patterns. According to our benchmark, in real-world usage interactive segmentation models may perform worse than it has been reported in the baseline benchmark, and most of the methods are not robust. We believe that RClicks is a significant step towards creating interactive segmentation methods that provide the best user experience in real-world cases.

Current benchmarks for assessing vision-language models (VLMs) often focus on their perception or problem-solving capabilities and neglect other critical aspects such as fairness, multilinguality, or toxicity. Furthermore, they differ in their evaluation procedures and the scope of the evaluation, making it difficult to compare models. To address these issues, we extend the HELM framework to VLMs to present the Holistic Evaluation of Vision Language Models (VHELM). VHELM aggregates various datasets to cover one or more of the 9 aspects: visual perception, knowledge, reasoning, bias, fairness, multilinguality, robustness, toxicity, and safety. In doing so, we produce a comprehensive, multi-dimensional view of the capabilities of the VLMs across these important factors. In addition, we standardize the standard inference parameters, methods of prompting, and evaluation metrics to enable fair comparisons across models. Our framework is designed to be lightweight and automatic so that evaluation runs are cheap and fast. Our initial run evaluates 22 VLMs on 21 existing datasets to provide a holistic snapshot of the models. We uncover new key findings, such as the fact that efficiency-focused models (e.g., Claude 3 Haiku or Gemini 1.5 Flash) perform significantly worse than their full models (e.g., Claude 3 Opus or Gemini 1.5 Pro) on the bias benchmark but not when evaluated on the other aspects. For transparency, we release the raw model generations and complete results on our website (https://crfm.stanford.edu/helm/vhelm/v2.0.1). VHELM is intended to be a living benchmark, and we hope to continue adding new datasets and models over time.