Related papers: Implicit Intelligence -- Evaluating Agents on What…
Prompt injection is the most critical vulnerability in deployed AI agents. Despite recent progress, we show that the prevailing defense paradigm (data-instruction separation) both fails to detect attacks that operate through contextual…
Recent work proposes using world models to generate controlled virtual environments in which AI agents can be tested before deployment to ensure their reliability and safety. However, accurate world models often have high computational…
Machine Learning (ML) has emerged as a powerful form of data modelling with widespread applicability beyond its roots in the design of autonomous agents. However, relatively little attention has been paid to the interaction between people…
The development of sophisticated artificial intelligence (AI) conversational agents based on large language models raises important questions about the relationship between human norms, values, and practices and AI design and performance.…
In the vision and language navigation task, the agent may encounter ambiguous situations that are hard to interpret by just relying on visual information and natural language instructions. We propose an interactive learning framework to…
Agents are a special kind of AI-based software in that they interact in complex environments and have increased potential for emergent behaviour. Explaining such emergent behaviour is key to deploying trustworthy AI, but the increasing…
Explanations for AI models in high-stakes domains like medicine often lack verifiability, which can hinder trust. To address this, we propose an interactive agent that produces explanations through an auditable sequence of actions. The…
The interpretation of implicit meanings is an integral aspect of human communication. However, this framework may not transfer to interactions with Large Language Models (LLMs). To investigate this, we introduce the task of Implicit…
Many real-world tasks involve a mixed-initiative setup, wherein humans and AI systems collaboratively perform a task. While significant work has been conducted towards enabling humans to specify, through language, exactly how an agent…
With the rapid improvement of LLMs' coding capabilities, the bottleneck of LLM-based automated software development is shifting from generating correct code to eliciting users' requirements. Despite growing interest, the interview…
The rise of AI conversational agents has broadened opportunities to enhance human capabilities across various domains. As these agents become more prevalent, it is crucial to investigate the impact of different affective abilities on their…
Large language models (LLMs) are increasingly leveraged to empower autonomous agents to simulate human beings in various fields of behavioral research. However, evaluating their capacity to navigate complex social interactions remains a…
AI design characteristics and human personality traits each impact the quality and outcomes of human-AI interactions. However, their relative and joint impacts are underexplored in imperfectly cooperative scenarios, where people and AI only…
Agentic frameworks are the software layer through which AI agents act in the world. Existing safety methods intervene on the model and therefore remain conditional on unverifiable properties of learned behavior. We introduce containment…
Active learning agents typically employ a query selection algorithm which solely considers the agent's learning objectives. However, this may be insufficient in more realistic human domains. This work uses imitation learning to enable an…
Most AI agents remain confined to an instrumental "command-execution" model, resulting in unequal, one-sided interactions. While recent works attempt to build relationships through hidden memory backends, these invisible processes often…
As AI agents integrate into enterprise applications, their evaluation demands benchmarks that reflect the complexity of real-world operations. Instead, existing benchmarks overemphasize open-domains such as code, use narrow accuracy…
An auditor instructs an AI assistant: "open each file individually using the Read tool -- no scripts, no agents." The AI replies "Yes" -- then issues a single batched call summarizing all fifty files at once. We call this the Compliance…
This paper introduces the concept of Language-Guided World Models (LWMs) -- probabilistic models that can simulate environments by reading texts. Agents equipped with these models provide humans with more extensive and efficient control,…
While we do not always use words, communicating what we want to an AI is a conversation -- with ourselves as well as with it, a recurring loop with optional steps depending on the complexity of the situation and our request. Any given…