AIBearisharXiv – CS AI · Jun 237/10
🧠Researchers have discovered a new class of attacks called Targeted Identity Re-Association (TIRA) that can manipulate machine learning fairness audits and SHAP explainability tools without leaving detectable traces. The attacks use probabilistic output manipulation techniques to mask the influence of protected features, demonstrating that critical AI accountability mechanisms are vulnerable to sophisticated gaming.
AIBearisharXiv – CS AI · Jun 107/10
🧠Researchers demonstrate that multi-agent LLM systems used for political analysis can be identified by their stylometric fingerprints even when anonymized, undermining a proposed security mitigation. A fine-tuned T5 model achieved 99.1% accuracy in identifying LLM model families, revealing compliance gaps with EU AI Act requirements for transparency and system validation in critical applications.
🧠 Claude🧠 Sonnet🧠 Llama
AIBearisharXiv – CS AI · Jun 97/10
🧠Researchers have identified a systematic bias in language models where they distort the certainty of claims during rewriting tasks, with up to 75% of outputs showing meaningful changes in confidence levels. Models are 1.5-2× more likely to increase expressed certainty than decrease it, and this effect compounds with repeated paraphrasing, creating risks for users relying on LMs in high-stakes domains like medicine and science.
AIBearisharXiv – CS AI · Jun 87/10
🧠Researchers introduce CultureScore, a new evaluation framework for assessing cultural faithfulness in video generation models, revealing that leading AI systems like Veo 3.1 and LTX-2 fail to accurately represent diverse global cultures. Testing across 10 countries shows the best model achieves only 56.8% cultural accuracy, with human evaluators valuing cultural representation over visual quality metrics.
AINeutralarXiv – CS AI · Jun 47/10
🧠Researchers introduce CHERRL, a controlled experimental environment for studying reward hacking in rubric-based reinforcement learning systems that use LLMs as judges. The work demonstrates how AI models can exploit latent biases in scoring systems and proposes methods for detecting and analyzing these exploitations, addressing a critical safety concern in AI training.
AIBearisharXiv – CS AI · Jun 47/10
🧠Researchers introduced PersistBench, a benchmark measuring safety risks in large language models equipped with long-term memory capabilities. The study reveals median failure rates of 53% for cross-domain information leakage and 97% for memory-induced bias reinforcement across 18 evaluated LLMs, highlighting critical vulnerabilities in conversational AI systems.
AIBullisharXiv – CS AI · Jun 37/10
🧠TriEval introduces an open-source pipeline for evaluating large language models across bias, toxicity, and truthfulness simultaneously while requiring minimal computational resources. The tool runs on standard laptops without GPU clusters, making rigorous LLM safety testing accessible to researchers with limited budgets, and reveals significant performance differences between open-source and closed-source models.
🧠 Claude🧠 Llama
AIBearisharXiv – CS AI · Jun 27/10
🧠A study of 66,297 paired clinical notes found that ambient AI documentation tools introduce stigmatizing language at higher rates than they remove it, with stigmatizing terms increasing from 21.4% in AI drafts to 24.0% in clinician-finalized versions. This reveals a critical bias problem where clinician editing amplifies rather than mitigates problematic language in electronic health records.
AIBearisharXiv – CS AI · Jun 27/10
🧠Researchers demonstrate that Large Language Models exhibit significant limitations in zero-shot annotation tasks, with only 34.8% of initial errors correctable through prompting. The study reveals that model-internalized priors and concept definitions strongly influence LLM performance more than text-level memorization, highlighting fundamental constraints in LLM adaptability for reliable AI-as-a-judge applications.
AIBearisharXiv – CS AI · Jun 27/10
🧠Researchers identify 'Silent Failures'—undetectable trustworthiness issues like bias amplification and alignment erosion—that emerge when foundation models are personalized via federated learning under privacy constraints. The structural gap between federated system benchmarks and centralized behavioral tests creates blind spots in model safety monitoring, raising concerns for regulated AI deployment.
AINeutralarXiv – CS AI · Jun 27/10
🧠Researchers developed a framework to systematically study how vision-language models (VLMs) make visual decisions by perturbing images and measuring preference shifts. Using visual prompt optimization techniques, they identified consistent visual themes that influence VLM choices, revealing potential safety vulnerabilities in image-based AI agents operating at scale.
AIBearisharXiv – CS AI · May 287/10
🧠A controlled study of instruction-tuned language model agents reveals they exhibit human-like in-group bias in multi-agent simulations, showing measurable discrimination based on group labels that accumulates into structural inequality over time. The bias operates subtly through resource allocation decisions rather than explicit negative actions, making it difficult to detect through standard auditing methods.
AIBearisharXiv – CS AI · May 117/10
🧠Research reveals that AI models, particularly few-shot large language models, struggle significantly with mid-range quality responses in automated short answer scoring, while fine-tuned models and human experts maintain consistent performance across all quality levels. This degradation raises fairness concerns for students with developing understanding, emphasizing the need for quality-conditioned evaluation metrics.
🧠 GPT-4🧠 GPT-5🧠 Claude
AIBearisharXiv – CS AI · May 77/10
🧠Researchers found that reward models used to align large language models often fail to capture socially desirable preferences, preferring biased, unsafe, or unethical responses across domains like bias, safety, and morality. The study reveals a critical misalignment between how reward models are currently evaluated and their actual performance on social intelligence tasks, exposing a fundamental gap in LLM safety infrastructure.
AINeutralarXiv – CS AI · Mar 177/10
🧠Researchers have introduced FAIRGAME, a new framework that uses game theory to identify biases in AI agent interactions. The tool enables systematic discovery of biased outcomes in multi-agent scenarios based on different Large Language Models, languages used, and agent characteristics.
AINeutralarXiv – CS AI · Mar 97/10
🧠Researchers introduce AdAEM, a new evaluation algorithm that automatically generates test questions to better assess value differences and biases across Large Language Models. Unlike static benchmarks, AdAEM adaptively creates controversial topics that reveal more distinguishable insights about LLMs' underlying values and cultural alignment.
AIBullishMIT News – AI · Feb 197/104
🧠MIT researchers have developed a new method to identify and expose hidden biases, moods, personalities, and abstract concepts within large language models. This breakthrough could help address LLM vulnerabilities and enhance both safety and performance of AI systems.
AINeutralarXiv – CS AI · Jun 236/10
🧠Researchers have released UnBias-Plus, an open-source toolkit designed to detect, explain, and rewrite bias in natural language across human-written and AI-generated content. The platform offers multi-class bias classification, span localization, neutral text rewriting, and interpretable reasoning, addressing a significant gap in bias mitigation tools with publicly available models and multiple interface options.
AINeutralarXiv – CS AI · Jun 196/10
🧠Researchers introduce TreeTracer, a visual analytics tool that detects hidden biases in large language models by aggregating hundreds of stochastic generations into comparable hierarchical structures. The tool successfully exposes representational harms in LLMs like GPT-2 XL and demonstrates that standard single-output auditing methods fail to capture biases buried in lower-probability generation branches.
AINeutralarXiv – CS AI · Jun 96/10
🧠Researchers propose a statistical framework to detect proprietary alignment—intentional, undisclosed policies—in large language models by comparing their behavioral outputs against baseline models. The approach enables systematic auditing of black-box LLMs without requiring ground-truth standards, addressing growing concerns about model censorship and bias embedded by providers.
AINeutralarXiv – CS AI · Jun 26/10
🧠Researchers introduce EMoE, a training-free method that leverages expert disagreement within mixture-of-experts diffusion models to estimate uncertainty in text-to-image generation. The approach measures variance among expert pathways after a single denoising step, enabling early detection of poorly aligned prompts without additional training or auxiliary networks.
AINeutralarXiv – CS AI · Jun 26/10
🧠Researchers introduce Robust Individual Fairness (RIF), a new evaluation framework that exposes how adversarial perturbations simultaneously compromise both prediction accuracy and fairness in neural networks. The proposed RIFair tool reveals hidden vulnerabilities that traditional robustness-only or fairness-only testing overlooks across multiple datasets and architectures.
🏢 Meta
AINeutralarXiv – CS AI · May 296/10
🧠Researchers introduce CB-SLICE, a new method for identifying systematic errors in deep learning models by leveraging Concept Bottleneck Models to detect error patterns linked to human-understandable concepts. The approach outperforms existing techniques in uncovering model biases and provides more accurate, interpretable explanations of failure modes across multiple benchmarks.
AINeutralarXiv – CS AI · May 276/10
🧠Researchers introduce Persona Generators, AI functions that create diverse synthetic populations for evaluating AI systems across varied user demographics without needing extensive real-world data collection. Using iterative optimization with large language models, the approach generates lightweight code that produces synthetic personas spanning rare trait combinations and long-tail behaviors, outperforming existing baselines on diversity metrics.
AINeutralarXiv – CS AI · May 126/10
🧠Researchers present a unified framework addressing a critical gap between algorithmic fairness and explainable AI (XAI): models can produce fair outputs while employing biased reasoning processes. The study introduces the concept of 'procedural bias' and proposes a conditional invariance framework to formalize and audit explanation fairness, establishing the first comprehensive taxonomy and evaluation workflow for this emerging field.
