AI Detection
Benchmarking AI Text Detectors Under Attack
A new benchmark evaluates AI-generated text detectors across model families, domains, and adversarial rewrites, highlighting how fragile authenticity tools can be outside narrow test settings.
research
research
SCE-LITE-HQ: Smoother Visual Counterfactuals via Diffusion
New research combines StyleGAN and diffusion models to generate high-quality visual counterfactual explanations, advancing explainable AI while revealing techniques applicable to synthetic media generation.
Voice AI
New Hierarchical Model Improves Real-Time Conversational AI Turn-
Researchers introduce a hierarchical end-of-turn detection model with primary speaker segmentation, advancing real-time conversational AI systems for more natural voice interactions.
AI ethics
Research Explores How Humans Assign Blame to AI Systems
New research investigates how people attribute causality to AI across scenarios of agency, misuse, and misalignment, with implications for accountability in synthetic media and deepfake governance.
Deepfake Detection
York University Team Advances Audio Deepfake Detection Methods
York University's forensic speech science team earns recognition at major Deepfake Detection Challenge, advancing techniques for identifying synthetic audio and voice cloning fraud.
Deepfake Detection
TMU Researchers Develop New Deepfake Detection Safeguards
Toronto Metropolitan University researchers are developing technical safeguards to combat deepfake deception, addressing the growing challenge of synthetic media authentication in digital spaces.
Blockchain
AI Meets Blockchain: Technical Framework for Decentralized Authen
New research explores how combining AI and blockchain creates robust systems for digital authenticity, content provenance, and decentralized verification in an era of synthetic media.
AI Safety
New Benchmark Measures AI Agents' Multi-Step Cyber Attack Abiliti
Researchers develop framework to measure how well AI agents can execute complex, multi-step cyber attacks, revealing critical insights for AI safety and security.
AI Agents
Counting AI Agents: New Framework Tackles Liability Questions
New research proposes frameworks for identifying and counting AI agents—a critical question as autonomous systems create content and take actions with real-world consequences.
Explainable AI
FAME Brings Formal Verification to Neural Network Explanations
New research introduces FAME, a framework using formal methods to generate mathematically guaranteed minimal explanations for neural network decisions, advancing AI interpretability.
super-resolution
Rank-Factorized Neural Bias Enables Scalable Super-Resolution
New research combines rank-factorized implicit neural bias with FlashAttention to scale super-resolution transformers efficiently, advancing high-quality image synthesis for AI-generated content.
AI Safety
LLM Safety Judges Are No Better Than Coin Flips, Study Finds
New research reveals LLM-based safety evaluators fail to reliably measure adversarial robustness, raising critical questions about automated AI safety testing methodologies.