Last Thanksgiving, Brian Israel found himself being asked the same question again and again. The general counsel at the AI lab Anthropic had been watching dumbfounded along with the rest of the tech world as, just two miles south of Anthropic's headquarters in San Francisco, its main competitor OpenAI seemed to be imploding. OpenAI's board had fired CEO Sam Altman, saying he had lost their confidence, in a move that seemed likely to tank the startup's 80 billion-plus valuation. The firing was only possible thanks to OpenAI's strange corporate structure, in which its directors have no fiduciary duty to increase profits for shareholders--a structure Altman himself had helped design so that OpenAI could build powerful AI insulated from perverse market incentives. To many, it appeared that plan had badly backfired.

As former President Donald Trump awaits a Manhattan jury's verdict, he can be forgiven for feeling that his criminal trial resembles a surreal "Alice in Wonderland" farce. He is left to peer through a "Looking-Glass" where everything is backward. The culprit for this hallucinatory nightmare is District Attorney Alvin Bragg who brought a bizarre case based on warped interpretations of law and distorted facts. It is now up to twelve jurors to wade through the lunacy in search of the illusive truth. Bragg's fractured case requires the jury to reach several distinct conclusions on issues that make little sense to begin with.

The United States Department of Justice is quietly prosecuting a novel Espionage Act case involving a drone, a Chinese national, and classified nuclear submarines. The case is such a rarity that it appears to be the first known prosecution under a World War II–era law that bans photographing vital military installations using aircraft, showing how new technologies are leading to fresh national security and First Amendment issues. "This is definitely not something that the law has addressed to any significant degree," Emily Berman, a law professor at the University of Houston who specializes in national security, tells WIRED. "There's definitely no reported cases." On January 5, 2024, Fengyun Shi flew to Virginia while on leave from his graduate studies at the University of Minnesota and rented a Tesla at the airport.

When you picture the tech industry, you probably think of things that don't exist in physical space, such as the apps and internet browser on your phone. But the infrastructure required to store all this information – the physical datacentres housed in business parks and city outskirts – consume massive amounts of energy. Despite its name, the infrastructure used by the "cloud" accounts for more global greenhouse emissions than commercial flights. In 2018, for instance, the 5bn YouTube hits for the viral song Despacito used the same amount of energy it would take to heat 40,000 US homes annually. This is a hugely environmentally destructive side to the tech industry.

The wide spread of Automatic Identification System (AIS) has motivated several maritime analytics operations. Vessel Location Forecasting (VLF) is one of the most critical operations for maritime awareness. However, accurate VLF is a challenging problem due to the complexity and dynamic nature of maritime traffic conditions. Furthermore, as privacy concerns and restrictions have grown, training data has become increasingly fragmented, resulting in dispersed databases of several isolated data silos among different organizations, which in turn decreases the quality of learning models. In this paper, we propose an efficient VLF solution based on LSTM neural networks, in two variants, namely Nautilus and FedNautilus for the centralized and the federated learning approach, respectively. We also demonstrate the superiority of the centralized approach with respect to current state of the art and discuss the advantages and disadvantages of the federated against the centralized approach.

Event Causality Identification (ECI) aims to detect whether there exists a causal relation between two events in a document. Existing studies adopt a kind of identifying after learning paradigm, where events' representations are first learned and then used for the identification. Furthermore, they mainly focus on the causality existence, but ignoring causal direction. In this paper, we take care of the causal direction and propose a new identifying while learning mode for the ECI task. We argue that a few causal relations can be easily identified with high confidence, and the directionality and structure of these identified causalities can be utilized to update events' representations for boosting next round of causality identification. To this end, this paper designs an *iterative learning and identifying framework*: In each iteration, we construct an event causality graph, on which events' causal structure representations are updated for boosting causal identification. Experiments on two public datasets show that our approach outperforms the state-of-the-art algorithms in both evaluations for causality existence identification and direction identification.

The task of persona-steered text generation requires large language models (LLMs) to generate text that reflects the distribution of views that an individual fitting a persona could have. People have multifaceted personas, but prior work on bias in LLM-generated opinions has only explored multiple-choice settings or one-dimensional personas. We define an incongruous persona as a persona with multiple traits where one trait makes its other traits less likely in human survey data, e.g. political liberals who support increased military spending. We find that LLMs are 9.7% less steerable towards incongruous personas than congruous ones, sometimes generating the stereotypical stance associated with its demographic rather than the target stance. Models that we evaluate that are fine-tuned with Reinforcement Learning from Human Feedback (RLHF) are more steerable, especially towards stances associated with political liberals and women, but present significantly less diverse views of personas. We also find variance in LLM steerability that cannot be predicted from multiple-choice opinion evaluation. Our results show the importance of evaluating models in open-ended text generation, as it can surface new LLM opinion biases. Moreover, such a setup can shed light on our ability to steer models toward a richer and more diverse range of viewpoints.

Underwater object tracking (UOT) is a foundational task for identifying and tracing submerged entities in underwater video sequences. However, current UOT datasets suffer from limitations in scale, diversity of target categories and scenarios covered, hindering the training and evaluation of modern tracking algorithms. To bridge this gap, we take the first step and introduce WebUOT-1M, \ie, the largest public UOT benchmark to date, sourced from complex and realistic underwater environments. It comprises 1.1 million frames across 1,500 video clips filtered from 408 target categories, largely surpassing previous UOT datasets, \eg, UVOT400. Through meticulous manual annotation and verification, we provide high-quality bounding boxes for underwater targets. Additionally, WebUOT-1M includes language prompts for video sequences, expanding its application areas, \eg, underwater vision-language tracking. Most existing trackers are tailored for open-air environments, leading to performance degradation when applied to UOT due to domain gaps. Retraining and fine-tuning these trackers are challenging due to sample imbalances and limited real-world underwater datasets. To tackle these challenges, we propose a novel omni-knowledge distillation framework based on WebUOT-1M, incorporating various strategies to guide the learning of the student Transformer. To the best of our knowledge, this framework is the first to effectively transfer open-air domain knowledge to the UOT model through knowledge distillation, as demonstrated by results on both existing UOT datasets and the newly proposed WebUOT-1M. Furthermore, we comprehensively evaluate WebUOT-1M using 30 deep trackers, showcasing its value as a benchmark for UOT research by presenting new challenges and opportunities for future studies. The complete dataset, codes and tracking results, will be made publicly available.

Large Language Models (LLMs) are typically harmless but remain vulnerable to carefully crafted prompts known as ``jailbreaks'', which can bypass protective measures and induce harmful behavior. Recent advancements in LLMs have incorporated moderation guardrails that can filter outputs, which trigger processing errors for certain malicious questions. Existing red-teaming benchmarks often neglect to include questions that trigger moderation guardrails, making it difficult to evaluate jailbreak effectiveness. To address this issue, we introduce JAMBench, a harmful behavior benchmark designed to trigger and evaluate moderation guardrails. JAMBench involves 160 manually crafted instructions covering four major risk categories at multiple severity levels. Furthermore, we propose a jailbreak method, JAM (Jailbreak Against Moderation), designed to attack moderation guardrails using jailbreak prefixes to bypass input-level filters and a fine-tuned shadow model functionally equivalent to the guardrail model to generate cipher characters to bypass output-level filters. Our extensive experiments on four LLMs demonstrate that JAM achieves higher jailbreak success ($\sim$ $\times$ 19.88) and lower filtered-out rates ($\sim$ $\times$ 1/6) than baselines.

Legal practice has witnessed a sharp rise in products incorporating artificial intelligence (AI). Such tools are designed to assist with a wide range of core legal tasks, from search and summarization of caselaw to document drafting. But the large language models used in these tools are prone to "hallucinate," or make up false information, making their use risky in high-stakes domains. Recently, certain legal research providers have touted methods such as retrieval-augmented generation (RAG) as "eliminating" (Casetext, 2023) or "avoid[ing]" hallucinations (Thomson Reuters, 2023), or guaranteeing "hallucination-free" legal citations (LexisNexis, 2023). Because of the closed nature of these systems, systematically assessing these claims is challenging. In this article, we design and report on the first preregistered empirical evaluation of AI-driven legal research tools. We demonstrate that the providers' claims are overstated. While hallucinations are reduced relative to general-purpose chatbots (GPT-4), we find that the AI research tools made by LexisNexis (Lexis+ AI) and Thomson Reuters (Westlaw AI-Assisted Research and Ask Practical Law AI) each hallucinate between 17% and 33% of the time. We also document substantial differences between systems in responsiveness and accuracy. Our article makes four key contributions. It is the first to assess and report the performance of RAG-based proprietary legal AI tools. Second, it introduces a comprehensive, preregistered dataset for identifying and understanding vulnerabilities in these systems. Third, it proposes a clear typology for differentiating between hallucinations and accurate legal responses. Last, it provides evidence to inform the responsibilities of legal professionals in supervising and verifying AI outputs, which remains a central open question for the responsible integration of AI into law.