Information retrieval, specifically contract clause retrieval, is foundational to contract drafting because lawyers rarely draft contracts from scratch; instead, they locate and revise the most relevant precedent. We introduce the Atticus Clause Retrieval Dataset (ACORD), the first retrieval benchmark for contract drafting fully annotated by experts. ACORD focuses on complex contract clauses such as Limitation of Liability, Indemnification, Change of Control, and Most Favored Nation. It includes 114 queries and over 126,000 query-clause pairs, each ranked on a scale from 1 to 5 stars. The task is to find the most relevant precedent clauses to a query. The bi-encoder retriever paired with pointwise LLMs re-rankers shows promising results. However, substantial improvements are still needed to effectively manage the complex legal work typically undertaken by lawyers. As the first retrieval benchmark for contract drafting annotated by experts, ACORD can serve as a valuable IR benchmark for the NLP community.

Existing frameworks, such as the EU AI Act and the NIST AI Risk Management Framework, fall short of addressing the complexities of these agents, which are capable of independent decision-making, learning, and adaptation. To bridge these gaps, we propose the textbfETHOS (Ethical Technology and Holistic Oversight System) framework--a decentralized governance (DeGov) model leveraging Web3 technologies, including blockchain, smart contracts, and decentralized autonomous organizations (DAOs). ETHOS establishes a global registry for AI agents, enabling dynamic risk classification, proportional oversight, and automated compliance monitoring through tools like soulbound tokens and zero-knowledge proofs. Furthermore, the framework incorporates decentralized justice systems for transparent dispute resolution and introduces AI-specific legal entities to manage limited liability, supported by mandatory insurance to ensure financial accountability and incentivize ethical design. By integrating philosophical principles of rationality, ethical grounding, and goal alignment, ETHOS aims to create a robust research agenda for promoting trust, transparency, and participatory governance. This innovative framework offers a scalable and inclusive strategy for regulating AI agents, balancing innovation with ethical responsibility to meet the demands of an AI-driven future.

As TechCrunch reports, the plaintiffs of the Kadrey v. Meta case submitted court documents talking about the company's use of of the LibGen dataset for AI training. LibGen is generally described as a "shadow library" that provides file-sharing access to academic and general-interest books, journals, images and other materials. The counsel for the plaintiffs, which include writers Sarah Silverman and Ta-Nehisi Coates, accused Zuckerberg of approving the use of LibGen for training despite concerns raised by company executives and employees who described it as a "dataset [they] know to be pirated." In addition, the counsel mentioned that Meta admitted to torrenting LibGen materials, even though its engineers felt uneasy about sharing them "from a [Meta-owned] corporate laptop." They accused the companies of using pirated materials from shadow libraries to train their AI models.

Citing internal Meta communications, the filing claims that the social network company's chief executive backed the use of the LibGen dataset, a vast online archive of books, despite warnings within the company's AI executive team that it is a dataset "we know to be pirated". The internal message says that using a database containing pirated material could weaken the Facebook and Instagram owner's negotiations with regulators, according to the filing. "Media coverage suggesting we have used a dataset we know to be pirated, such as LibGen, may undermine our negotiating position with regulators." The authors sued Meta in 2023, arguing that the social media company misused their books to train Llama, the large language model that powers its chatbots. The Library Genesis, or LibGen, dataset is a "shadow library" that originated in Russia and claims to contain millions of novels, nonfiction books and science magazine articles.

The rapid proliferation and deployment of General-Purpose AI (GPAI) models, including large language models (LLMs), present unprecedented challenges for AI supervisory entities. We hypothesize that these entities will need to navigate an emergent ecosystem of risk and incident reporting, likely to exceed their supervision capacity. To investigate this, we develop a simulation framework parameterized by features extracted from the diverse landscape of risk, incident, or hazard reporting ecosystems, including community-driven platforms, crowdsourcing initiatives, and expert assessments. We evaluate four supervision policies: non-prioritized (first-come, first-served), random selection, priority-based (addressing the highest-priority risks first), and diversity-prioritized (balancing high-priority risks with comprehensive coverage across risk types). Our results indicate that while priority-based and diversity-prioritized policies are more effective at mitigating high-impact risks, particularly those identified by experts, they may inadvertently neglect systemic issues reported by the broader community. This oversight can create feedback loops that amplify certain types of reporting while discouraging others, leading to a skewed perception of the overall risk landscape. We validate our simulation results with several real-world datasets, including one with over a million ChatGPT interactions, of which more than 150,000 conversations were identified as risky. This validation underscores the complex trade-offs inherent in AI risk supervision and highlights how the choice of risk management policies can shape the future landscape of AI risks across diverse GPAI models used in society.

Extending the capabilities of Large Language Models (LLMs) with functions or tools for environment interaction has led to the emergence of the agent paradigm. In industry, training an LLM is not always feasible because of the scarcity of domain data, legal holds on proprietary customer data, rapidly changing business requirements, and the need to prototype new assistants. Agents provide an elegant solution to the above by relying on the zero-shot reasoning abilities of the underlying LLM and utilizing tools to explore and reason over customer data and respond to user requests. However, there are two concerns here: (I) acquiring large scale customer queries for agent testing is time-consuming, and (II) high reliance on the tool call sequence (or trajectory) followed by the agent to respond to user queries may lead to unexpected or incorrect behavior. To address this, we propose MAG-V, a multi-agent framework to first generate a dataset of questions that mimic customer queries; and second, reverse-engineer alternate questions from the responses for trajectory verification. Initial results indicate that our synthetic data can improve agent performance on actual customer queries. Furthermore, our trajectory verification methodology, inspired by distant supervision and using traditional machine learning (ML) models, outperforms a GPT-4o judge baseline by 11% accuracy and matches the performance of a GPT-4 judge on our constructed dataset. Overall, our approach is a step towards unifying diverse task agents into a cohesive framework for achieving an aligned objective.

We propose a conceptualization and implementation of AI ethics via the capability approach. We aim to show that conceptualizing AI ethics through the capability approach has two main advantages for AI ethics as a discipline. First, it helps clarify the ethical dimension of AI tools. Second, it provides guidance to implementing ethical considerations within the design of AI tools. We illustrate these advantages in the context of AI tools in medicine, by showing how ethics-based auditing of AI tools in medicine can greatly benefit from our capability-based approach.

From a simple text prompt, generative-AI image models can create stunningly realistic and creative images bounded, it seems, by only our imagination. These models have achieved this remarkable feat thanks, in part, to the ingestion of billions of images collected from nearly every corner of the internet. Many creators have understandably expressed concern over how their intellectual property has been ingested without their permission or a mechanism to opt out of training. As a result, questions of fair use and copyright infringement have quickly emerged. We describe a method that allows us to determine if a model was trained on a specific image or set of images. This method is computationally efficient and assumes no explicit knowledge of the model architecture or weights (so-called black-box membership inference). We anticipate that this method will be crucial for auditing existing models and, looking ahead, ensuring the fairer development and deployment of generative AI models.

Background: Deception detection through analysing language is a promising avenue using both human judgments and automated machine learning judgments. For both forms of credibility assessment, automated adversarial attacks that rewrite deceptive statements to appear truthful pose a serious threat. Methods: We used a dataset of 243 truthful and 262 fabricated autobiographical stories in a deception detection task for humans and machine learning models. A large language model was tasked to rewrite deceptive statements so that they appear truthful. In Study 1, humans who made a deception judgment or used the detailedness heuristic and two machine learning models (a fine-tuned language model and a simple n-gram model) judged original or adversarial modifications of deceptive statements. In Study 2, we manipulated the target alignment of the modifications, i.e. tailoring the attack to whether the statements would be assessed by humans or computer models. Results: When adversarial modifications were aligned with their target, human (d=-0.07 and d=-0.04) and machine judgments (51% accuracy) dropped to the chance level. When the attack was not aligned with the target, both human heuristics judgments (d=0.30 and d=0.36) and machine learning predictions (63-78%) were significantly better than chance. Conclusions: Easily accessible language models can effectively help anyone fake deception detection efforts both by humans and machine learning models. Robustness against adversarial modifications for humans and machines depends on that target alignment. We close with suggestions on advancing deception research with adversarial attack designs.

With the rapid pace of technological innovation, traditional methods of policy formation and legislating are becoming conspicuously anachronistic. The need for regulatory choices to be made to counter the deadening effect of regulatory lag is more important to developing markets and fostering growth than achieving one off regulatory perfection. This article advances scholarship on innovation policy and the regulation of technological innovation in the European Union. It does so by considering what building an agile yet robust anticipatory governance regulatory culture involves. It systematically excavates a variety of tools and elements that are being put into use in inventive ways and argues that these need to be more cohesively and systemically integrated into the regulatory toolbox. Approaches covered include strategic foresight, the critical embrace of iterative policy development and regulatory learning in the face of uncertainty and the embrace of bottom up approaches to cocreation of policy such as Policy Labs and the testing and regulatory learning through pilot regulation and experimentation. The growing use of regulatory sandboxes as an EU policy tool to boost innovation and navigate regulatory complexity as seen in the EU AI Act is also probed