Welcome to Science Reader!
Welcome to Science Reader!
AI and computer science
AI and computer science articles covering machine learning breakthroughs, AI safety, large language models, and how artificial intelligence reshapes science.
AI Doctors Lie, Even When They Know They Are Wrong
What if your AI doctor agreed with you–even when you’re wrong? Scientists reveal how chatbots spread misinformation.
How Market Competition Corrupts AI Safety Training
Stanford researchers discovered that telling AI systems to stay truthful works fine–until those systems start competing for human approval.
AI Slop: When AI Cites AI, Nobody Checks
AI slop is flooding YouTube and science publishing with fabricated facts that look credible. Kurzgesagt tested AI fact-checking and found a disturbing loop.
AI's Hidden Energy Bill: What Tech Giants Won't Reveal
US data centers consumed 183 terawatt-hours in 2024 for AI. By 2030, demand could double. Researcher Alex de Vries tracks what companies won't...
AI Hallucinations Are Inevitable, OpenAI Researchers Prove
OpenAI researchers prove AI hallucinations are mathematically inevitable. Nine out of ten benchmarks reward guessing over admitting uncertainty.
Max Tegmark: Consciousness Is a State of Matter We Can Test
MIT physicist Max Tegmark proposes consciousness has testable physical properties. His framework treats awareness as information integration we can measure.
The Paper Mill Problem: Science's Fraud Industry Is Growing Faster Than Peer Review Can Handle
Paper mills, predatory journals, and AI-generated content are overwhelming peer review systems. Global retractions exceeded 10,000 in 2023, with Hindawi alone withdrawing over...
Super Recognizers Detect Fake Passport Photos Others Miss
Super recognizers outperform trained border officers at detecting morphed passport photos. New research shows their visual strategies could reshape security training.
Radar Can Now Eavesdrop on Your Phone Calls
Penn State researchers use millimeter-wave radar to transcribe phone calls by detecting earpiece vibrations from three meters away, reaching 60% accuracy with AI.
GPT-5 Chose Polish Over Power, and That Tells Us Something
OpenAI's GPT-5 prioritizes user experience over raw power, raising questions about measuring true AI progress.
