机器之心报道 编辑：杜伟、Panda 这届 ICLR 的烦心事还没有结束。 最近一段时间，对于 ICLR 2026 来说，真可谓是一波未平、一波又起。先是第三方机构对审稿意见的系统性统计发现，其中 有 21% 完全由 AI 生成 ；后有 OpenReview 评审大开盒 ，波及到了 ICLR 2026 超过 10000 篇投稿。 今天，ICLR 2026 的审稿又被揭开一块遮羞布。事情是这样的： AI 生成内容检测平台 GPTZero 扫描了 300 篇 投稿论文，发现其中有 50 篇在论文引用上至少包含 一处明显的幻觉内容。 甚至有些幻觉引用还非常离谱，达到了匪夷所思的程度，就好像投稿者完全不检查一样。比如下面 GPTZero CTO 和联创 Alex Cui 在 X 分享的这个例子，投稿者给 出的引用链接竟然是默认示例链接 example.com ！ 而在下面的例子中，作者名单就只是一串大写字母。 更令人担忧的是， 这些存在幻觉内容的投稿已经经过了 3-5 名领域专家的同行评审，但他们中的绝大多数都未能识别出这些虚假的引用。 这意味着，如果这些投稿没有其他外部干预，就可能会被 ICLR 会议接收。部分投稿 ...

Core Insights - The article discusses a significant paper led by Fei-Fei Li that establishes a clear framework for the emerging field of Agent AI, outlining its capabilities and potential applications [5][6][9] - The paper presents a comprehensive cognitive architecture for Agent AI, consisting of five core modules: Environment and Perception, Cognition, Action, Learning, and Memory, which together form a dynamic and iterative closed-loop system [11][12][18] Summary by Sections Agent AI Framework - The new Agent AI paradigm is not merely a combination of existing technologies but represents a forward-thinking approach to the development of Artificial General Intelligence (AGI) [12] - The framework integrates various technological strands, including dialogue models, visual-language models, and reinforcement learning, into a unified perspective on multimodal agents [9][12] Core Modules of Agent AI - **Environment and Perception**: This module allows agents to actively perceive information from the physical or virtual world, incorporating task planning and skill observation [13] - **Cognition**: Defined as the processing center of the agent, this module utilizes large language models (LLMs) and visual-language models (VLMs) to interpret sensory information and develop strategies [14] - **Action**: This module generates specific operational commands based on cognitive decisions, enabling interaction with both physical and virtual environments [15] - **Learning**: Emphasizes the agent's ability to continuously learn and evolve through various mechanisms, including reinforcement learning and imitation learning [16] - **Memory**: Unlike traditional models, this module provides a structured and persistent memory system that allows agents to leverage past experiences for future tasks [17][18] Role of Large Models - Large foundational models, particularly LLMs and VLMs, serve as the cognitive backbone of Agent AI, enabling agents to perform complex tasks with minimal predefined rules [20] - The paper highlights the challenge of "hallucination," where models generate inaccurate content, and proposes environmental interaction as a solution to mitigate this issue [21] Ethical and Regulatory Considerations - The article stresses the importance of inclusivity and ethical considerations in the design of Agent AI, advocating for diverse training data and bias detection mechanisms [22] - It also addresses the need for clear regulations and frameworks to ensure data privacy and security, especially in sensitive applications [22] Application Potential - **Gaming**: Agent AI can revolutionize non-player character (NPC) behavior, allowing for dynamic interactions and personalized experiences in gaming environments [25][26] - **Robotics**: Agents can autonomously plan and execute complex physical tasks based on natural language commands, enhancing user interaction with robots [28] - **Healthcare**: Agent AI can assist in preliminary medical consultations and patient monitoring, significantly improving healthcare delivery, especially in resource-limited settings [30][32] Future Directions - The article acknowledges that Agent AI is still in its early stages and faces challenges in achieving deep integration across various modalities and domains [33] - It emphasizes the need for standardized evaluation metrics to assess agent intelligence and guide future research [33]

Core Viewpoint - The article discusses the challenges posed by AI hallucinations, particularly their impact on the application of AI technologies across various sectors, emphasizing the need for effective governance and mitigation strategies [1][2]. Group 1: Understanding AI Hallucinations - AI hallucinations are primarily defined as discrepancies between generated content and reality, often due to training design flaws, insufficient data, and architectural biases [2][3]. - There are three main types of AI hallucinations: factual hallucinations (creation of false events or knowledge), fidelity hallucinations (inconsistencies in long text), and cross-modal inconsistencies (discrepancies in multi-modal outputs) [3][4]. - The phenomenon of AI hallucinations is viewed as an inherent aspect of AI evolution, suggesting that some level of creative freedom in generation is necessary for AI's capability breakthroughs [4][6]. Group 2: Implications of AI Hallucinations - The dangers of AI hallucinations vary significantly depending on the application context; for instance, using AI for casual conversation poses less risk than in critical fields like healthcare or law [7][8]. - AI hallucinations can lead to severe consequences in high-stakes environments, such as legal proceedings where fabricated references can disrupt judicial processes [9][10]. - The potential for AI-generated content to pollute the internet and exacerbate the spread of misinformation is a growing concern, particularly as these inaccuracies may be used as training data for future models [9][10]. Group 3: Mitigation Strategies - Experts suggest that employing high-quality training datasets, including both native and synthetic data, is essential to reduce the occurrence of AI hallucinations [11][12]. - Implementing verification mechanisms, such as knowledge graphs and causal inference models, can enhance the reliability of AI outputs [12][13]. - Regulatory measures, including mandatory labeling of AI-generated content, are being established to improve transparency and accountability in AI applications [14].