Artificial Intelligence & Ethics

Global: Policy Challenges for General-Purpose AIPolicymakers are confronting new challenges as general-purpose artificial intelligence (GPAI) gains prominence, prompting a reassessment of regulatory frameworks that were originally crafted for task‑specific AI…

Global: Oracle Interventions Reveal Critical Skills for Multi‑Turn AI AgentsA team of artificial‑intelligence researchers announced a new study on Jan. 26, 2026, via an arXiv preprint, that proposes an oracle counterfactual framework…

Global: Adaptive Framework Proposed to Strengthen Large Language Model Safety A new adaptive safety framework for large language models (LLMs) called SafeThinker was introduced in a paper submitted to arXiv…

Global: Activation Probes Provide Low-Cost Monitoring for High-Stakes LLM InteractionsOn January 23, 2026, a team of seven researchers announced a new method for flagging potentially harmful language model outputs. The…

Global: New Differential Privacy Structure Secures Cross-Attention in Generative AIResearchers have unveiled a novel data structure that enforces differential privacy for cross‑attention mechanisms in large generative models, according to a…

Global: Adaptive Spatial Goodness Encoding Enhances Forward-Forward Training for Convolutional NetworksResearchers have introduced Adaptive Spatial Goodness Encoding (ASGE), a new training framework that replaces backpropagation with the Forward-Forward (FF) algorithm…

Global: Federated Learning Framework Enables Privacy-Preserving Training of Multimodal ModelsResearchers have introduced FedUMM, a federated learning framework designed for unified multimodal models (UMMs), in a preprint posted on arXiv. The…

Global: Research Finds Transformer Models May Skew Sentiment NeutralityOn January 21, 2026, researcher Prasanna Kumar submitted a preprint to arXiv titled “The Dark Side of AI Transformers: Sentiment Polarization &…

Global: Advanced Retrieval-Augmented AI Shows Near-Zero Fabrication in Legal Task EvaluationResearchers led by Alex Dantart released a study on Jan. 21, 2026 that evaluates the reliability of large language models…

Global: New Framework Captures Multiplicity and Tail Risks in Generative AI EvaluationA team of AI researchers has introduced a unified framework aimed at improving how generative models are assessed, according…