LLM safety and ethical alignment are widely discussed, but the impact of content moderation on user satisfaction remains underexplored. In particular, little is known about how users respond when models refuse to answer a prompt-one of the primary mechanisms used to enforce ethical boundaries in LLMs. We address this gap by analyzing nearly 50,000 model comparisons from Chatbot Arena, a platform where users indicate their preferred LLM response in pairwise matchups, providing a large-scale setting for studying real-world user preferences. Using a novel RoBERTa-based refusal classifier fine-tuned on a hand-labeled dataset, we distinguish between refusals due to ethical concerns and technical limitations. Our results reveal a substantial refusal penalty: ethical refusals yield significantly lower win rates than both technical refusals and standard responses, indicating that users are especially dissatisfied when models decline a task for ethical reasons. However, this penalty is not uniform. Refusals receive more favorable evaluations when the underlying prompt is highly sensitive (e.g., involving illegal content), and when the refusal is phrased in a detailed and contextually aligned manner. These findings underscore a core tension in LLM design: safety-aligned behaviors may conflict with user expectations, calling for more adaptive moderation strategies that account for context and presentation.
As large language models (LLMs) are increasingly deployed in diverse applications, including chatbot assistants and code generation, aligning their behavior with safety and ethical standards has become paramount. However, jailbreak attacks, which exploit vulnerabilities to elicit unintended or harmful outputs, threaten LLMs' safety significantly. In this paper, we introduce Layer-AdvPatcher, a novel methodology designed to defend against jailbreak attacks by utilizing an unlearning strategy to patch specific layers within LLMs through self-augmented datasets. Our insight is that certain layer(s), tend to produce affirmative tokens when faced with harmful prompts. By identifying these layers and adversarially exposing them to generate more harmful data, one can understand their inherent and diverse vulnerabilities to attacks. With these exposures, we then "unlearn" these issues, reducing the impact of affirmative tokens and hence minimizing jailbreak risks while keeping the model's responses to safe queries intact. We conduct extensive experiments on two models, four benchmark datasets, and multiple state-of-the-art jailbreak attacks to demonstrate the efficacy of our approach. Results indicate that our framework reduces the harmfulness and attack success rate of jailbreak attacks without compromising utility for benign queries compared to recent defense methods. Our code is publicly available at: https://github.com/oyy2000/LayerAdvPatcher
Question answering systems are rapidly advancing, but their opaque nature may impact user trust. We explored trust through an anti-monitoring framework, where trust is predicted to be correlated with presence of citations and inversely related to checking citations. We tested this hypothesis with a live question-answering experiment that presented text responses generated using a commercial Chatbot along with varying citations (zero, one, or five), both relevant and random, and recorded if participants checked the citations and their self-reported trust in the generated responses. We found a significant increase in trust when citations were present, a result that held true even when the citations were random; we also found a significant decrease in trust when participants checked the citations. These results highlight the importance of citations in enhancing trust in AI-generated content.
One of the long-standing aspirations in conversational AI is to allow them to autonomously take initiatives in conversations, i.e., being proactive. This is especially challenging for multi-party conversations. Prior NLP research focused mainly on predicting the next speaker from contexts like preceding conversations. In this paper, we demonstrate the limitations of such methods and rethink what it means for AI to be proactive in multi-party, human-AI conversations. We propose that just like humans, rather than merely reacting to turn-taking cues, a proactive AI formulates its own inner thoughts during a conversation, and seeks the right moment to contribute. Through a formative study with 24 participants and inspiration from linguistics and cognitive psychology, we introduce the Inner Thoughts framework. Our framework equips AI with a continuous, covert train of thoughts in parallel to the overt communication process, which enables it to proactively engage by modeling its intrinsic motivation to express these thoughts. We instantiated this framework into two real-time systems: an AI playground web app and a chatbot. Through a technical evaluation and user studies with human participants, our framework significantly surpasses existing baselines on aspects like anthropomorphism, coherence, intelligence, and turn-taking appropriateness.
A chatbot is an intelligent software application that automates conversations and engages users in natural language through messaging platforms. Leveraging artificial intelligence (AI), chatbots serve various functions, including customer service, information gathering, and casual conversation. Existing virtual assistant chatbots, such as ChatGPT and Gemini, demonstrate the potential of AI in Natural Language Processing (NLP). However, many current solutions rely on predefined APIs, which can result in vendor lock-in and high costs. To address these challenges, this work proposes a chatbot developed using a Sequence-to-Sequence (Seq2Seq) model with an encoder-decoder architecture that incorporates attention mechanisms and Long Short-Term Memory (LSTM) cells. By avoiding predefined APIs, this approach ensures flexibility and cost-effectiveness. The chatbot is trained, validated, and tested on a dataset specifically curated for the tourism sector in Draa-Tafilalet, Morocco. Key evaluation findings indicate that the proposed Seq2Seq model-based chatbot achieved high accuracies: approximately 99.58% in training, 98.03% in validation, and 94.12% in testing. These results demonstrate the chatbot's effectiveness in providing relevant and coherent responses within the tourism domain, highlighting the potential of specialized AI applications to enhance user experience and satisfaction in niche markets.
This work investigates the causal mechanism behind the effect of chatbot personification and persuasion strategies on users' perceptions and donation likelihood. In a 2 (personified vs. non-personified chatbot) x 2 (emotional vs. logical persuasion strategy) between-subjects experiment (N=76), participants engaged with a chatbot that represented a non-profit charitable organization. The results suggest that interaction with a personified chatbot evokes perceived anthropomorphism; however, it does not elicit greater willingness to donate. In fact, we found that commonly used anthropomorphic features, like name and narrative, led to negative attitudes toward an AI agent in the donation context. Our results showcase a preference for non-personified chatbots paired with logical persuasion appeal, emphasizing the significance of consistency in chatbot interaction, mirroring human-human engagement. We discuss the importance of moving from exploring the common scenario of a chatbot with machine identity vs. a chatbot with human identity in light of the recent regulations of AI systems.
Large Language Models (LLMs) are emerging as powerful tools in healthcare, particularly for complex, domain-specific tasks. This study describes the development and evaluation of the PErioperative AI CHatbot (PEACH), a secure LLM-based system integrated with local perioperative guidelines to support preoperative clinical decision-making. PEACH was embedded with 35 institutional perioperative protocols in the secure Claude 3.5 Sonet LLM framework within Pair Chat (developed by Singapore Government) and tested in a silent deployment with real-world data. Accuracy, safety, and usability were assessed. Deviations and hallucinations were categorized based on potential harm, and user feedback was evaluated using the Technology Acceptance Model (TAM). Updates were made after the initial silent deployment to amend one protocol. In 240 real-world clinical iterations, PEACH achieved a first-generation accuracy of 97.5% (78/80) and an overall accuracy of 96.7% (232/240) across three iterations. The updated PEACH demonstrated improved accuracy of 97.9% (235/240), with a statistically significant difference from the null hypothesis of 95% accuracy (p = 0.018, 95% CI: 0.952-0.991). Minimal hallucinations and deviations were observed (both 1/240 and 2/240, respectively). Clinicians reported that PEACH expedited decisions in 95% of cases, and inter-rater reliability ranged from kappa 0.772-0.893 within PEACH and 0.610-0.784 among attendings. PEACH is an accurate, adaptable tool that enhances consistency and efficiency in perioperative decision-making. Future research should explore its scalability across specialties and its impact on clinical outcomes.
Beyond mere formality, small talk plays a pivotal role in social dynamics, serving as a verbal handshake for building rapport and understanding. For conversational AI and social robots, the ability to engage in small talk enhances their perceived sociability, leading to more comfortable and natural user interactions. In this study, we evaluate the capacity of current Large Language Models (LLMs) to drive the small talk of a social robot and identify key areas for improvement. We introduce a novel method that autonomously generates feedback and ensures LLM-generated responses align with small talk conventions. Through several evaluations -- involving chatbot interactions and human-robot interactions -- we demonstrate the system's effectiveness in guiding LLM-generated responses toward realistic, human-like, and natural small-talk exchanges.
Large language models (LLMs) have transformed natural language processing, with frameworks like Chatbot Arena providing pioneering platforms for evaluating these models. By facilitating millions of pairwise comparisons based on human judgments, Chatbot Arena has become a cornerstone in LLM evaluation, offering rich datasets for ranking models in open-ended conversational tasks. Building upon this foundation, we propose a statistical framework that incorporates key advancements to address specific challenges in pairwise comparison analysis. First, we introduce a factored tie model that enhances the ability to handle ties -- an integral aspect of human-judged comparisons -- significantly improving the model's fit to observed data. Second, we extend the framework to model covariance between competitors, enabling deeper insights into performance relationships and facilitating intuitive groupings into performance tiers. Third, we resolve optimization challenges arising from parameter non-uniqueness by introducing novel constraints, ensuring stable and interpretable parameter estimation. Through rigorous evaluation and extensive experimentation, our framework demonstrates substantial improvements over existing methods in modeling pairwise comparison data. To support reproducibility and practical adoption, we release leaderbot, an open-source Python package implementing our models and analyses.
The rise of LLMs has deflected a growing portion of human-computer interactions towards LLM-based chatbots. The remarkable abilities of these models allow users to interact using long, diverse natural language text covering a wide range of topics and styles. Phrasing these messages is a time and effort consuming task, calling for an autocomplete solution to assist users. We introduce the task of chatbot interaction autocomplete. We present ChaI-TeA: CHat InTEraction Autocomplete; An autcomplete evaluation framework for LLM-based chatbot interactions. The framework includes a formal definition of the task, coupled with suitable datasets and metrics. We use the framework to evaluate After formally defining the task along with suitable datasets and metrics, we test 9 models on the defined auto completion task, finding that while current off-the-shelf models perform fairly, there is still much room for improvement, mainly in ranking of the generated suggestions. We provide insights for practitioners working on this task and open new research directions for researchers in the field. We release our framework to serve as a foundation for future research.
This pilot study presents the development of the InfoTech Assistant, a domain-specific, multimodal chatbot engineered to address queries in bridge evaluation and infrastructure technology. By integrating web data scraping, large language models (LLMs), and Retrieval-Augmented Generation (RAG), the InfoTech Assistant provides accurate and contextually relevant responses. Data, including textual descriptions and images, are sourced from publicly available documents on the InfoTechnology website and organized in JSON format to facilitate efficient querying. The architecture of the system includes an HTML-based interface and a Flask back end connected to the Llama 3.1 model via LLM Studio. Evaluation results show approximately 95 percent accuracy on domain-specific tasks, with high similarity scores confirming the quality of response matching. This RAG-enhanced setup enables the InfoTech Assistant to handle complex, multimodal queries, offering both textual and visual information in its responses. The InfoTech Assistant demonstrates strong potential as a dependable tool for infrastructure professionals, delivering high accuracy and relevance in its domain-specific outputs.
As artificial intelligence (AI) becomes increasingly central to various fields, there is a growing need to equip K-12 students with AI literacy skills that extend beyond computer science. This paper explores the integration of a Project-Based Learning (PBL) AI toolkit into diverse subject areas, aimed at helping educators teach AI concepts more effectively. Through interviews and co-design sessions with K-12 teachers, we examined current AI literacy levels and how teachers adapt AI tools like the AI Art Lab, AI Music Studio, and AI Chatbot into their course designs. While teachers appreciated the potential of AI tools to foster creativity and critical thinking, they also expressed concerns about the accuracy, trustworthiness, and ethical implications of AI-generated content. Our findings reveal the challenges teachers face, including limited resources, varying student and instructor skill levels, and the need for scalable, adaptable AI tools. This research contributes insights that can inform the development of AI curricula tailored to diverse educational contexts.
Chatbot research is advancing with the growing importance of chatbots in fields that require human interactions, such as customer support and mental health care. Despite these advancements, chatbots still face significant challenges in understanding subtle nuances and managing long conversation histories. To address these issues, our study introduces a dual approach: firstly, we employ Emotional Preference Optimization (EPO) to train chatbots not only with correct responses but also with counter-emotional responses-those that are contextually similar but emotionally divergent. This training enables the model to discern fine nuance distinctions between correct and counter-emotional responses, thereby enhancing the quality of its responses. Secondly, we introduce MambaCompressor to effectively compress and manage extensive conversation histories, significantly reducing time and memory complexities while improving the chatbot's contextual understanding. Our comprehensive experiments across multiple datasets demonstrate that our model significantly outperforms existing models in generating empathetic responses and efficiently managing lengthy dialogues.
Anxiety and depression are the most common mental health issues worldwide, affecting a non-negligible part of the population. Accordingly, stakeholders, including governments' health systems, are developing new strategies to promote early detection and prevention from a holistic perspective (i.e., addressing several disorders simultaneously). In this work, an entirely novel system for the multi-label classification of anxiety and depression is proposed. The input data consists of dialogues from user interactions with an assistant chatbot. Another relevant contribution lies in using Large Language Models (LLMs) for feature extraction, provided the complexity and variability of language. The combination of LLMs, given their high capability for language understanding, and Machine Learning (ML) models, provided their contextual knowledge about the classification problem thanks to the labeled data, constitute a promising approach towards mental health assessment. To promote the solution's trustworthiness, reliability, and accountability, explainability descriptions of the model's decision are provided in a graphical dashboard. Experimental results on a real dataset attain 90 % accuracy, improving those in the prior literature. The ultimate objective is to contribute in an accessible and scalable way before formal treatment occurs in the healthcare systems.
Problem: Pancreas radiological imaging is challenging due to the small size, blurred boundaries, and variability of shape and position of the organ among patients. Goal: In this work we present MiniGPT-Pancreas, a Multimodal Large Language Model (MLLM), as an interactive chatbot to support clinicians in pancreas cancer diagnosis by integrating visual and textual information. Methods: MiniGPT-v2, a general-purpose MLLM, was fine-tuned in a cascaded way for pancreas detection, tumor classification, and tumor detection with multimodal prompts combining questions and computed tomography scans from the National Institute of Health (NIH), and Medical Segmentation Decathlon (MSD) datasets. The AbdomenCT-1k dataset was used to detect the liver, spleen, kidney, and pancreas. Results: MiniGPT-Pancreas achieved an Intersection over Union (IoU) of 0.595 and 0.550 for the detection of pancreas on NIH and MSD datasets, respectively. For the pancreas cancer classification task on the MSD dataset, accuracy, precision, and recall were 0.876, 0.874, and 0.878, respectively. When evaluating MiniGPT-Pancreas on the AbdomenCT-1k dataset for multi-organ detection, the IoU was 0.8399 for the liver, 0.722 for the kidney, 0.705 for the spleen, and 0.497 for the pancreas. For the pancreas tumor detection task, the IoU score was 0.168 on the MSD dataset. Conclusions: MiniGPT-Pancreas represents a promising solution to support clinicians in the classification of pancreas images with pancreas tumors. Future research is needed to improve the score on the detection task, especially for pancreas tumors.
Diagnostic errors in healthcare persist as a critical challenge, with increasing numbers of patients turning to online resources for health information. While AI-powered healthcare chatbots show promise, there exists no standardized and scalable framework for evaluating their diagnostic capabilities. This study introduces a scalable benchmarking methodology for assessing health AI systems and demonstrates its application through August, an AI-driven conversational chatbot. Our methodology employs 400 validated clinical vignettes across 14 medical specialties, using AI-powered patient actors to simulate realistic clinical interactions. In systematic testing, August achieved a top-one diagnostic accuracy of 81.8% (327/400 cases) and a top-two accuracy of 85.0% (340/400 cases), significantly outperforming traditional symptom checkers. The system demonstrated 95.8% accuracy in specialist referrals and required 47% fewer questions compared to conventional symptom checkers (mean 16 vs 29 questions), while maintaining empathetic dialogue throughout consultations. These findings demonstrate the potential of AI chatbots to enhance healthcare delivery, though implementation challenges remain regarding real-world validation and integration of objective clinical data. This research provides a reproducible framework for evaluating healthcare AI systems, contributing to the responsible development and deployment of AI in clinical settings.
The deployment of large language models (LLMs) in diverse applications requires a thorough understanding of their decision-making strategies and behavioral patterns. As a supplement to a recent study on the behavioral Turing test, this paper presents a comprehensive analysis of five leading LLM-based chatbot families as they navigate a series of behavioral economics games. By benchmarking these AI chatbots, we aim to uncover and document both common and distinct behavioral patterns across a range of scenarios. The findings provide valuable insights into the strategic preferences of each LLM, highlighting potential implications for their deployment in critical decision-making roles.
Evaluating the quality of recommender systems is critical for algorithm design and optimization. Most evaluation methods are computed based on offline metrics for quick algorithm evolution, since online experiments are usually risky and time-consuming. However, offline evaluation usually cannot fully reflect users' preference for the outcome of different recommendation algorithms, and the results may not be consistent with online A/B test. Moreover, many offline metrics such as AUC do not offer sufficient information for comparing the subtle differences between two competitive recommender systems in different aspects, which may lead to substantial performance differences in long-term online serving. Fortunately, due to the strong commonsense knowledge and role-play capability of large language models (LLMs), it is possible to obtain simulated user feedback on offline recommendation results. Motivated by the idea of LLM Chatbot Arena, in this paper we present the idea of RecSys Arena, where the recommendation results given by two different recommender systems in each session are evaluated by an LLM judger to obtain fine-grained evaluation feedback. More specifically, for each sample we use LLM to generate a user profile description based on user behavior history or off-the-shelf profile features, which is used to guide LLM to play the role of this user and evaluate the relative preference for two recommendation results generated by different models. Through extensive experiments on two recommendation datasets in different scenarios, we demonstrate that many different LLMs not only provide general evaluation results that are highly consistent with canonical offline metrics, but also provide rich insight in many subjective aspects. Moreover, it can better distinguish different algorithms with comparable performance in terms of AUC and nDCG.
In recent years, we are seeing considerable interest in conversational agents with the rise of large language models (LLMs). Although they offer considerable advantages, LLMs also present significant risks, such as hallucination, which hinder their widespread deployment in industry. Moreover, low-resource languages such as African ones are still underrepresented in these systems limiting their performance in these languages. In this paper, we illustrate a more classical approach based on modular architectures of Task-oriented Dialog Systems (ToDS) offering better control over outputs. We propose a chatbot generation engine based on the Rasa framework and a robust methodology for projecting annotations onto the Wolof language using an in-house machine translation system. After evaluating a generated chatbot trained on the Amazon Massive dataset, our Wolof Intent Classifier performs similarly to the one obtained for French, which is a resource-rich language. We also show that this approach is extensible to other low-resource languages, thanks to the intent classifier's language-agnostic pipeline, simplifying the design of chatbots in these languages.
Eating disorders (ED) are complex mental health conditions that require long-term management and support. Recent advancements in large language model (LLM)-based chatbots offer the potential to assist individuals in receiving immediate support. Yet, concerns remain about their reliability and safety in sensitive contexts such as ED. We explore the opportunities and potential harms of using LLM-based chatbots for ED recovery. We observe the interactions between 26 participants with ED and an LLM-based chatbot, WellnessBot, designed to support ED recovery, over 10 days. We discovered that our participants have felt empowered in recovery by discussing ED-related stories with the chatbot, which served as a personal yet social avenue. However, we also identified harmful chatbot responses, especially concerning individuals with ED, that went unnoticed partly due to participants' unquestioning trust in the chatbot's reliability. Based on these findings, we provide design implications for safe and effective LLM-based interventions in ED management.