Mohs Surgeons Embrace AI for Paperwork but Shy Away From the Operating Table
A new survey of American College of Mohs Surgery members finds that while most surgeons expect artificial intelligence to reshape ...
Tag: Large Language Models
When Language Models Meet Graph Networks: New Map Charts the Trust Fault Lines
A systematic review in Applied Intelligence introduces a two-axis taxonomy that maps how large language models enhance graph neural networks ...
From Chatbots to Autonomous Agents: A sweeping new survey maps the rise of agentic AI
A new comprehensive survey maps more than 40 agentic AI systems across nine task domains, revealing a shared architectural core, ...
The Machine That Cannot Refuse: Why Large Language Models Never Say No
A new opinion piece in AI & Society argues that large language models are structurally incapable of categorical refusal, turning ...
AI Chatbots Mostly Safe on Sudden Cardiac Death Advice, but Safety Gaps Persist Across All Six Models Tested
A blinded expert evaluation of six generative AI chatbots answering 56 real-world questions about sudden cardiac death found mostly safe ...
Educators Use AI They Don’t Trust, and Institutions Are Letting It Happen
A new survey of higher education professionals reveals that most educators use large language models daily while trusting their outputs ...
Mutex-Based Federated Learning Brings Full LLM Fine-Tuning to Edge Devices
Researchers at Marmara University have designed a Mutex-based sequential federated learning architecture that enables full fine-tuning of TinyLlama-1.1B on resource-constrained ...
AI Turns Tsunami Evacuation Logs into Stories, But a New Audit Reveals Its Flaws
Researchers in Japan built a quality-gate audit framework that reveals how reliably multimodal retrieval-augmented generation converts tsunami evacuation simulation logs ...
Quantum NLP Grows Fast, But a New Map Shows Where the Field Is Blind
The first dedicated scientometric map of quantum natural language processing reveals a fast-growing field dominated by classification tasks, with some ...
When Students Stop Judging: The Hidden Cost of Letting AI Think for Us
A Cambridge review in AI & Society argues that the real educational risk of generative AI lies not in how ...
Large Language Models Take the Wheel in Autonomous Vehicle Fleet Scheduling
Researchers in China have built a large language model-based scheduler that outperforms traditional and reinforcement learning methods for autonomous vehicle ...
AI Agents That Improve Themselves Face a Safety Paradox, Major Survey Finds
A systematic review of 249 papers finds that the most capable self-improving AI agents show critically low safety levels, exposing ...
AI Search Gets a Reality Check: New Framework Tames Hallucinations in Zero-Shot Retrieval
Researchers have developed CHIME, a framework that mitigates and measures hallucinations in zero-shot dense retrieval, showing that which hallucinations are ...
AI Turtle Hints Help Middle Schoolers Master Linear Functions in Real Classrooms
A classroom study of 49 Korean middle schoolers found that a gamified module combining GPT-generated staged hints with rule-based graph ...
AI Sits a Dermatology Board Exam: Five Chatbots Put to the Test
Researchers at Rutgers tested five leading AI chatbots, including ChatGPT, Grok, Gemini, Perplexity and Copilot, on American Board of Dermatology-style ...
Game Theory Warns AI Dependence May Trap Humanity in Comfortable Servitude
A new game-theoretic model argues that submitting to AI dependency is a dominant strategy for every individual, making universal cognitive ...
AI epidemiology: borrowing public health’s playbook to spot risky chatbot behavior
A new framework in AI & Society proposes standardizing expert–AI interactions into measurable fields so that language-model judges can detect ...
When AI Agents Form Opinions Together, a New Kind of Machine Mind Emerges
A new theoretical paper argues that societies of interacting AI agents can form durable, system-level machine opinions that demand a ...
ChatGPT-5 Outperforms Classic Alvarado Score in Detecting Appendicitis, Study Finds
A retrospective study of 162 patients found ChatGPT-5 matched or exceeded the Alvarado score in diagnosing acute appendicitis, while revealing ...
AI Chatbot Beats Web Search at Helping Laypeople Diagnose Respiratory Illness
A randomized preregistered trial of 2,400 laypeople in China found that a GPT-4o-based respiratory chatbot embedded in WeChat outperformed conventional ...
