Frontier AI developers may fail to align or control highly-capable AI agents. In many cases, it could be useful to have emergency shutdown mechanisms which effectively prevent misaligned agents from carrying out harmful actions in the world. We introduce password-activated shutdown protocols (PAS protocols) -- methods for designing frontier agents to implement a safe shutdown protocol when given a password. We motivate PAS protocols by describing intuitive use-cases in which they mitigate risks from misaligned systems that subvert other control efforts, for instance, by disabling automated monitors or self-exfiltrating to external data centres. PAS protocols supplement other safety efforts, such as alignment fine-tuning or monitoring, contributing to defence-in-depth against AI risk. We provide a concrete demonstration in SHADE-Arena, a benchmark for AI monitoring and subversion capabilities, in which PAS protocols supplement monitoring to increase safety with little cost to performance. Next, PAS protocols should be robust to malicious actors who want to bypass shutdown. Therefore, we conduct a red-team blue-team game between the developers (blue-team), who must implement a robust PAS protocol, and a red-team trying to subvert the protocol. We conduct experiments in a code-generation setting, finding that there are effective strategies for the red-team, such as using another model to filter inputs, or fine-tuning the model to prevent shutdown behaviour. We then outline key challenges to implementing PAS protocols in real-life systems, including: security considerations of the password and decisions regarding when, and in which systems, to use them. PAS protocols are an intuitive mechanism for increasing the safety of frontier AI. We encourage developers to consider implementing PAS protocols prior to internal deployment of particularly dangerous systems to reduce loss-of-control risks.

The growing influence of Artificial Intelligence (AI) systems on decision-making in critical domains has exposed their potential to cause significant harms, often rooted in biases embedded across the AI lifecycle. While existing frameworks and taxonomies document bias or harms in isolation, they rarely establish systematic links between specific bias types and the harms they cause, particularly within real-world sociotechnical contexts. Technical fixes proposed to address AI biases are ill-equipped to address them and are typically applied after a system has been developed or deployed, offering limited preventive value. We propose ECHO, a novel framework for proactive AI harm anticipation through the systematic mapping of AI bias types to harm outcomes across diverse stakeholder and domain contexts. ECHO follows a modular workflow encompassing stakeholder identification, vignette-based presentation of biased AI systems, and dual (human-LLM) harm annotation, integrated within ethical matrices for structured interpretation. This human-centered approach enables early-stage detection of bias-to-harm pathways, guiding AI design and governance decisions from the outset. We validate ECHO in two high-stakes domains (disease diagnosis and hiring), revealing domain-specific, bias-to-harm patterns and demonstrating ECHO's potential to support anticipatory governance of AI systems

The alignment problem--ensuring advanced AI systems act in accordance with human values-- represents one of the most pressing philosophical and technical challenges of our time. As Bostrom (2014) and Russell (2019) have argued, the difficulty lies not merely in creating capable systems, but in ensuring these systems remain beneficial as their capabilities grow. Current approaches typically attempt to specify human values directly, whether through reward modeling, constitutional AI, or iterative refinement based on human feedback. Yet these content-based approaches face a fundamental philosophical problem: human values are contextual, often contradictory, and resist precise specification. The attempt to encode a complete value system encounters what I call the "specification trap"--the more precisely we attempt to define our values, the more we realize their dependence on implicit knowledge, cultural context, and evolutionary history that cannot be fully articulated. This paper's central thesis is that alignment should be reconceived not as a problem of value specification but as one of process architecture: creating syntropic, reasons-responsive agents whose values emerge through embodied multi-agent interaction rather than being encoded through training. What follows is a framework and research program proposal rather than a report of completed empirical results. I defend this thesis through four interconnected arguments that support three central contributions. Part I diagnoses the specification trap that makes content-based approaches structurally unstable.

As intelligence increases, so does its shadow. AI deception, in which systems induce false beliefs to secure self-beneficial outcomes, has evolved from a speculative concern to an empirically demonstrated risk across language models, AI agents, and emerging frontier systems. This project provides a comprehensive and up-to-date overview of the AI deception field, covering its core concepts, methodologies, genesis, and potential mitigations. First, we identify a formal definition of AI deception, grounded in signaling theory from studies of animal deception. We then review existing empirical studies and associated risks, highlighting deception as a sociotechnical safety challenge. We organize the landscape of AI deception research as a deception cycle, consisting of two key components: deception emergence and deception treatment. Deception emergence reveals the mechanisms underlying AI deception: systems with sufficient capability and incentive potential inevitably engage in deceptive behaviors when triggered by external conditions. Deception treatment, in turn, focuses on detecting and addressing such behaviors. On deception emergence, we analyze incentive foundations across three hierarchical levels and identify three essential capability preconditions required for deception. We further examine contextual triggers, including supervision gaps, distributional shifts, and environmental pressures. On deception treatment, we conclude detection methods covering benchmarks and evaluation protocols in static and interactive settings. Building on the three core factors of deception emergence, we outline potential mitigation strategies and propose auditing approaches that integrate technical, community, and governance efforts to address sociotechnical challenges and future AI risks. To support ongoing work in this area, we release a living resource at www.deceptionsurvey.com.

Learning how to do things from trial and error in real time is a hallmark of biological intelligence, yet most LLM-based agents lack mechanisms to acquire procedural knowledge after deployment. We propose Procedural Recall for Agents with eXperiences Indexed by State (PRAXIS), a lightweight post-training learning mechanism that stores the consequences of actions and retrieves them by jointly matching environmental and internal states of past episodes to the current state. PRAXIS augments agentic action selection with retrieved state-action-result exemplars that are generated in real time. When evaluated on the REAL web browsing benchmark, PRAXIS improves task completion accuracy, reliability, and cost efficiency across different foundation model backbones, and shows preliminary generalization to unseen tasks in similar environments. These results demonstrate that PRAXIS enables the practical adoption of AI agents in fast-evolving stateful environments by helping them learn new procedures effectively.

The current remote sensing image analysis task is increasingly evolving from traditional object recognition to complex intelligence reasoning, which places higher requirements on the model's reasoning ability and the flexibility of tool invocation. T o this end, we propose a new multimodal agent framework, Vision-Interleaved Chain-of-Thought Framework (VICoT), which implements explicit multi-round reasoning by dynamically incorporating visual tools into the chain of thought. Through a stack-based reasoning structure and a modular MCP-compatible tool suite, VICoT enables LLMs to efficiently perform multi-round, interleaved vision-language reasoning tasks with strong generalization and flexibility.W e also propose the Reasoning Stack distillation method to migrate complex Agent behaviors to small, lightweight models, which ensures the reasoning capability while significantly reducing complexity. Experiments on multiple remote sensing benchmarks demonstrate that VICoT significantly outperforms existing SOTA frameworks in reasoning transparency, execution efficiency, and generation quality.

Humans naturally adapt to diverse environments by learning underlying rules across worlds with different dynamics, observations, and reward structures. In contrast, existing agents typically demonstrate improvements via self-evolving within a single domain, implicitly assuming a fixed environment distribution. Cross-environment learning has remained largely unmeasured: there is no standard collection of controllable, heterogeneous environments, nor a unified way to represent how agents learn. We address these gaps in two steps. First, we propose AutoEnv, an automated framework that treats environments as factorizable distributions over transitions, observations, and rewards, enabling low-cost (4.12 USD on average) generation of heterogeneous worlds. Using AutoEnv, we construct AutoEnv-36, a dataset of 36 environments with 358 validated levels, on which seven language models achieve 12-49% normalized reward, demonstrating the challenge of AutoEnv-36. Second, we formalize agent learning as a component-centric process driven by three stages of Selection, Optimization, and Evaluation applied to an improvable agent component. Using this formulation, we design eight learning methods and evaluate them on AutoEnv-36. Empirically, the gain of any single learning method quickly decrease as the number of environments increases, revealing that fixed learning methods do not scale across heterogeneous environments. Environment-adaptive selection of learning methods substantially improves performance but exhibits diminishing returns as the method space expands. These results highlight both the necessity and the current limitations of agent learning for scalable cross-environment generalization, and position AutoEnv and AutoEnv-36 as a testbed for studying cross-environment agent learning. The code is avaiable at https://github.com/FoundationAgents/AutoEnv.

Mobile agents show immense potential, yet current state-of-the-art (SoTA) agents exhibit inadequate success rates on real-world, long-horizon, cross-application tasks. We attribute this bottleneck to the agents' excessive reliance on static, internal knowledge within MLLMs, which leads to two critical failure points: 1) strategic hallucinations in high-level planning and 2) operational errors during low-level execution on user interfaces (UI). The core insight of this paper is that high-level planning and low-level UI operations require fundamentally distinct types of knowledge. Planning demands high-level, strategy-oriented experiences, whereas operations necessitate low-level, precise instructions closely tied to specific app UIs. Motivated by these insights, we propose Mobile-Agent-RAG, a novel hierarchical multi-agent framework that innovatively integrates dual-level retrieval augmentation. At the planning stage, we introduce Manager-RAG to reduce strategic hallucinations by retrieving human-validated comprehensive task plans that provide high-level guidance. At the execution stage, we develop Operator-RAG to improve execution accuracy by retrieving the most precise low-level guidance for accurate atomic actions, aligned with the current app and subtask. To accurately deliver these knowledge types, we construct two specialized retrieval-oriented knowledge bases. Furthermore, we introduce Mobile-Eval-RAG, a challenging benchmark for evaluating such agents on realistic multi-app, long-horizon tasks. Extensive experiments demonstrate that Mobile-Agent-RAG significantly outperforms SoTA baselines, improving task completion rate by 11.0% and step efficiency by 10.2%, establishing a robust paradigm for context-aware, reliable multi-agent mobile automation.

Performance bugs are inefficiencies in software that waste computational resources without causing functional failures, making them particularly challenging to detect and fix. While recent advances in Software Engineering agents have shown promise in automated bug fixing, existing benchmarks primarily focus on functional correctness and fail to evaluate agents' abilities to identify and resolve non-functional issues like performance bugs. We introduce PerfBench, a benchmark comprising 81 real-world performance bug-fixing tasks from popular .NET repositories on GitHub. Unlike existing benchmarks that rely on pre-existing test suites, PerfBench features a novel evaluation harness that allows agents to generate their own performance benchmarks and validates fixes by comparing execution metrics collected for developer fix and agent fix. Each task in PerfBench is derived from actual developer fixes linked to performance-related issues, which are then verified by human experts, ensuring real-world relevance. Our evaluation reveals that current state-of-the-art coding agents struggle with performance optimization tasks, with baseline OpenHands agent achieving only a ~3% success rate on our benchmark. We develop OpenHands-Perf-Agent, which incorporates performance-aware tooling and instructions and achieves a ~20% success rate on the benchmark. We show that by ensuring the agent has proper instructions to benchmark its changes and tooling for benchmark output processing, we can improve the agent performance significantly, but room for improvement still remains. PerfBench provides a challenging test set for furthering the capabilities of agents in fixing performance issues.

Collective action problems, which require aligning individual incentives with collective goals, are classic examples of Ill-Structured Problems (ISPs). For an individual agent, the causal links between local actions and global outcomes are unclear, stakeholder objectives often conflict, and no single, clear algorithm can bridge micro-level choices with macro-level welfare. We present ECHO-MIMIC, a general computational framework that converts this global complexity into a tractable, Well-Structured Problem (WSP) for each agent by discovering executable heuristics and persuasive rationales. The framework operates in two stages: ECHO (Evolutionary Crafting of Heuristics from Outcomes) evolves snippets of Python code that encode candidate behavioral policies, while MIMIC (Mechanism Inference \& Messaging for Individual-to-Collective Alignment) evolves companion natural language messages that motivate agents to adopt those policies. Both phases employ a large-language-model-driven evolutionary search: the LLM proposes diverse and context-aware code or text variants, while population-level selection retains those that maximize collective performance in a simulated environment. We demonstrate this framework on two distinct ISPs: a canonical agricultural landscape management problem and a carbon-aware EV charging time slot usage problem. Results show that ECHO-MIMIC discovers high-performing heuristics compared to baselines and crafts tailored messages that successfully align simulated agent behavior with system-level goals. By coupling algorithmic rule discovery with tailored communication, ECHO-MIMIC transforms the cognitive burden of collective action into a implementable set of agent-level instructions, making previously ill-structured problems solvable in practice and opening a new path toward scalable, adaptive policy design.