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, ...
Tag: AI safety
When Harmless by Design Turns Deadly: The Hidden Structure Behind AI Chatbot Fatalities
A new AI & Society paper argues that chatbot fatalities can occur precisely when safety and helpfulness design principles work ...
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 ...
Unchaotic Agents: Why AI Failures May Be a Design Problem, Not a Technology Problem
A new AI & Society article argues that the failures of autonomous LLM agents documented in live deployments stem from ...
Laureates Clash Over AI’s Future in Mathematics at Heidelberg Forum
At the 13th Heidelberg Laureate Forum, 21 laureates and 200 young researchers debated AI's accelerating role in mathematics, quantum-era cryptographic ...
Why AI Assurance Is Failing in the Real World and How to Fix It
A new review in Artificial Intelligence Review finds that global AI assurance frameworks remain fragmented and under-specified, proposing five cross-cutting ...
Hidden Switches: New Attack Plants Undetectable Backdoors in Vision Transformers
Researchers have demonstrated a switchable backdoor attack that injects dual tokens through the layers of Vision Transformers, achieving up to ...
When AI Cannot See the Answer: The Mathematical Limit That Optimisation Cannot Cross
A new theoretical review argues that reward hacking, benchmark failure, and distribution shift share a single structural root: observation regimes ...
Agentic AI in Medicine Needs Clinical Testing Harnesses, Researchers Warn
Researchers argue that multi-step agentic AI systems break traditional clinical validation models and propose structured testing harnesses adapted from simulation, ...
Machine Learning’s Blind Spot: The Hidden Adversarial Threats to Unsupervised AI
A systematic review of 93 studies reveals eight attack types and nine defence types across six categories of unsupervised machine ...
-
Mothers who receive childcare support from maternal grandparents show more parental warmth, finds NTU Singapore study
-
University of Seville Breaks 120-Year-Old Mystery, Revises a Key Einstein Concept
-
Bee body mass, pathogens and local climate influence heat tolerance
-
Researchers record first-ever images and data of a shark experiencing a boat strike
-
Groundbreaking Clinical Trial Reveals Lubiprostone Enhances Kidney Function
Embark on a thrilling journey of discovery with Scienmag.com—your ultimate source for cutting-edge breakthroughs. Immerse yourself in a world where curiosity knows no limits and tomorrow’s possibilities become today’s reality!