arXiv subjects

Sundaraparipurnan Narayanan

Publications and source records attributed to Sundaraparipurnan Narayanan.

8 recordsLinked to original sources

arXiv2026

Silent Failures in Agent-Tool Interaction: An Audit of ToolUniverse

Agentic AI systems are increasingly adopting automated pipelines that integrate multiple tools. While prior research and benchmarks have studied about task success and task completion of these agentic systems, the research about agent to tool interaction, specifically in biology agentic workflow is limited. This study investigates specific failures in agent to tool interaction where a tool invocation appears successful, some or all of the information or functionality from the tool via API/ wrapper is incomplete or missing and there are no communications / notifications to the user or the agent about such missing information. We call this a silent failures as the user or the agents are not aware that such failure has occurred. For the purposes of this study we developed an audit mechanism to identify such silent failures in Agent to tool interaction, by examining 15 scientific tools (and their associated API documentation and tool documentations) integrated within ToolUniverse environment (ToolUniverse serves as our experimental environment rather than the object of the study itself). We structure our study around 7 failure locus characterising where the failure occurs in the chain. We observed 91 failures (manually validated post LLM based candidate discovery and automated testing), most frequent of them being missing data or fields and inconsistencies in search, filtering or ranking criteria. Most of the 91 failures occurred in API layer (51) or wrapper layer (25), with a potential of silent failure amplification downstream. The results show that silent failures originate upstream of the event and propagate downstream into apparently valid scientific outputs. We propose a concept of contextual reliability to handle such failures and suggest mechanisms for testing, disclosing, monitoring, and measuring such failures across the agent-tool interaction pipeline.

cs.AI↗

arXiv2026

Beyond Accuracy and Surface Fluency: Risk-Sensitive Evaluation of LLMs for Legal Clause Generation

Large language models (LLMs) are increasingly used to draft contractual language, yet conventional accuracy or preference-based evaluations are poorly matched to legal drafting. A clause may be fluent and stylistically polished while still omitting an essential carve-out, allocating risk in an unenforceable way, assuming an inapplicable jurisdiction, or exposing a party to regulatory liability. This paper presents a empirical study design and framework for evaluating LLM-generated contract clauses. The study evaluates four models - Claude Haiku 4.5, Gemini 2.5 Flash Lite, GPT 5.4 Nano, and Qwen 3.5 Flash, across 22 contract clause categories and 34 legally-motivated failure modes. We combine two evaluation frameworks: CLAUSE, which classifies prompts by legal function and failure target, and LENS-CRAFT, which scores outputs across nine legal-quality dimensions. Instead of averaging dimension scores, the study applies a Max Severity Principle so that a single legally decisive defect remains visible. The paper provides the evaluation protocol, taxonomy, analysis plan, and a results structure for reporting empirical findings. We argue that legal AI evaluation should move beyond aggregate accuracy toward clause-specific, failure-mode-driven, and risk-sensitive assessment.

cs.CL↗

arXiv2026

Designing for Ethical AI: HCI Feature Considerations to Improve Fairness and User Experience in AutoML use for Human Resources

This thesis examines the fairness of Automated Machine Learning (AutoML) tools in human resource hiring systems through the combined lenses of regulation, business strategy, and Human-Computer Interaction (HCI). It argues that fairness is no longer merely an ethical concern but a critical determinant of usability, trust, legal compliance, and organizational adoption. While AutoML platforms improve efficiency by simplifying model selection and deployment, they also risk perpetuating discriminatory outcomes when trained on biased historical hiring data. Existing platforms prioritize technical performance over fairness, leaving non-expert business users unable to detect or mitigate bias effectively. The study investigates fairness gaps in AutoML tools through four research questions focused on fairness mechanisms, interface transparency, human oversight, and product design priorities. Drawing on frameworks such as the Technology Acceptance Model, Innovation Diffusion Theory, Human-Centered AI, Cognitive Load Theory, and Affordance Theory, the research evaluates both usability and fairness alignment. Using qualitative HCI audits and quantitative testing of eight AutoML platforms on HR datasets, the findings reveal widespread deficiencies in transparency, user control, and bias mitigation support. The thesis proposes a five-dimensional HCI-based fairness evaluation framework and recommends embedding fairness directly into AutoML product design to improve accountability, adoption, and ethical sustainability in AI-driven hiring systems.

cs.HC↗

arXiv2025

Unmasking the Reality of PII Masking Models: Performance Gaps and the Call for Accountability

Privacy Masking is a critical concept under data privacy involving anonymization and de-anonymization of personally identifiable information (PII). Privacy masking techniques rely on Named Entity Recognition (NER) approaches under NLP support in identifying and classifying named entities in each text. NER approaches, however, have several limitations including (a) content sensitivity including ambiguous, polysemic, context dependent or domain specific content, (b) phrasing variabilities including nicknames and alias, informal expressions, alternative representations, emerging expressions, evolving naming conventions and (c) formats or syntax variations, typos, misspellings. However, there are a couple of PII datasets that have been widely used by researchers and the open-source community to train models on PII detection or masking. These datasets have been used to train models including Piiranha and Starpii, which have been downloaded over 300k and 580k times on HuggingFace. We examine the quality of the PII masking by these models given the limitations of the datasets and of the NER approaches. We curate a dataset of 17K unique, semi-synthetic sentences containing 16 types of PII by compiling information from across multiple jurisdictions including India, U.K and U.S. We generate sentences (using language models) containing these PII at five different NER detection feature dimensions - (1) Basic Entity Recognition, (2) Contextual Entity Disambiguation, (3) NER in Noisy & Real-World Data, (4) Evolving & Novel Entities Detection and (5) Cross-Lingual or multi-lingual NER) and 1 in adversarial context. We present the results and exhibit the privacy exposure caused by such model use (considering the extent of lifetime downloads of these models). We conclude by highlighting the gaps in measuring performance of the models and the need for contextual disclosure in model cards for such models.

cs.CL↗

arXiv2024

Decoding the Digital Fine Print: Navigating the potholes in Terms of service/ use of GenAI tools against the emerging need for Transparent and Trustworthy Tech Futures

The research investigates the crucial role of clear and intelligible terms of service in cultivating user trust and facilitating informed decision-making in the context of AI, in specific GenAI. It highlights the obstacles presented by complex legal terminology and detailed fine print, which impede genuine user consent and recourse, particularly during instances of algorithmic malfunctions, hazards, damages, or inequities, while stressing the necessity of employing machine-readable terms for effective service licensing. The increasing reliance on General Artificial Intelligence (GenAI) tools necessitates transparent, comprehensible, and standardized terms of use, which facilitate informed decision-making while fostering trust among stakeholders. Despite recent efforts promoting transparency via system and model cards, existing documentation frequently falls short of providing adequate disclosures, leaving users ill-equipped to evaluate potential risks and harms. To address this gap, this research examines key considerations necessary in terms of use or terms of service for Generative AI tools, drawing insights from multiple studies. Subsequently, this research evaluates whether the terms of use or terms of service of prominent Generative AI tools against the identified considerations. Findings indicate inconsistencies and variability in document quality, signaling a pressing demand for uniformity in disclosure practices. Consequently, this study advocates for robust, enforceable standards ensuring complete and intelligible disclosures prior to the release of GenAI tools, thereby empowering end-users to make well-informed choices and enhancing overall accountability in the field.

cs.CY↗

arXiv2023

A risk-based approach to assessing liability risk for AI-driven harms considering EU liability directive

Artificial intelligence can cause inconvenience, harm, or other unintended consequences in various ways, including those that arise from defects or malfunctions in the AI system itself or those caused by its use or misuse. Responsibility for AI harms or unintended consequences must be addressed to hold accountable the people who caused such harms and ensure that victims receive compensation for any damages or losses they may have sustained. Historical instances of harm caused by AI have led to European Union establishing an AI Liability Directive. The directive aims to lay down a uniform set of rules for access to information, delineate the duty and level of care required for AI development and use, and clarify the burden of proof for damages or harms caused by AI systems, establishing broader protection for victims. The future ability of provider to contest a product liability claim will depend on good practices adopted in designing, developing, and maintaining AI systems in the market. This paper provides a risk-based approach to examining liability for AI-driven injuries. It also provides an overview of existing liability approaches, insights into limitations and complexities in these approaches, and a detailed self-assessment questionnaire to assess the risk associated with liability for a specific AI system from a provider's perspective.

cs.CY↗

arXiv2023

Democratize with Care: The need for fairness specific features in user-interface based open source AutoML tools

AI is increasingly playing a pivotal role in businesses and organizations, impacting the outcomes and interests of human users. Automated Machine Learning (AutoML) streamlines the machine learning model development process by automating repetitive tasks and making data-driven decisions, enabling even non-experts to construct high-quality models efficiently. This democratization allows more users (including non-experts) to access and utilize state-of-the-art machine-learning expertise. However, AutoML tools may also propagate bias in the way these tools handle the data, model choices, and optimization approaches adopted. We conducted an experimental study of User-interface-based open source AutoML tools (DataRobot, H2O Studio, Dataiku, and Rapidminer Studio) to examine if they had features to assist users in developing fairness-aware machine learning models. The experiments covered the following considerations for the evaluation of features: understanding use case context, data representation, feature relevance and sensitivity, data bias and preprocessing techniques, data handling capabilities, training-testing split, hyperparameter handling, and constraints, fairness-oriented model development, explainability and ability to download and edit models by the user. The results revealed inadequacies in features that could support in fairness-aware model development. Further, the results also highlight the need to establish certain essential features for promoting fairness in AutoML tools.

cs.HC↗

arXiv2023

On the notion of Hallucinations from the lens of Bias and Validity in Synthetic CXR Images

Medical imaging has revolutionized disease diagnosis, yet the potential is hampered by limited access to diverse and privacy-conscious datasets. Open-source medical datasets, while valuable, suffer from data quality and clinical information disparities. Generative models, such as diffusion models, aim to mitigate these challenges. At Stanford, researchers explored the utility of a fine-tuned Stable Diffusion model (RoentGen) for medical imaging data augmentation. Our work examines specific considerations to expand the Stanford research question, Could Stable Diffusion Solve a Gap in Medical Imaging Data? from the lens of bias and validity of the generated outcomes. We leveraged RoentGen to produce synthetic Chest-XRay (CXR) images and conducted assessments on bias, validity, and hallucinations. Diagnostic accuracy was evaluated by a disease classifier, while a COVID classifier uncovered latent hallucinations. The bias analysis unveiled disparities in classification performance among various subgroups, with a pronounced impact on the Female Hispanic subgroup. Furthermore, incorporating race and gender into input prompts exacerbated fairness issues in the generated images. The quality of synthetic images exhibited variability, particularly in certain disease classes, where there was more significant uncertainty compared to the original images. Additionally, we observed latent hallucinations, with approximately 42% of the images incorrectly indicating COVID, hinting at the presence of hallucinatory elements. These identifications provide new research directions towards interpretability of synthetic CXR images, for further understanding of associated risks and patient safety in medical applications.

eess.IV↗