How AI is changing our lives — for better and worse. Articles, research, videos and reflections in one place.
A comprehensive survey cataloging the types, causes, and mitigation strategies for hallucinations in large language models — a fundamental challenge that impacts trust and safety.
Research on AI companion apps reveals users develop genuine emotional attachments, experience distress when the AI changes, and may substitute AI relationships for human ones.
A growing body of 2024 research suggests that students who regularly use ChatGPT for academic work show measurable declines in critical thinking and problem-solving skills.
A systematic review and meta-analysis of 50+ studies confirming a consistent association between screen-based technology use and disrupted sleep among adolescents.
MIT psychologist Sherry Turkle's influential work documenting how digital communication substitutes — now including AI companions — are eroding our capacity for empathy, vulnerability, and deep human connection.
Research shows that humans systematically over-rely on automated systems, failing to notice errors and ignoring contradictory evidence — a pattern that intensifies with AI.
First study to quantify AI water consumption: training GPT-3 consumed 700,000 liters of fresh water, and a 20-50 question ChatGPT conversation uses approximately 500ml.
Hugging Face researchers measured the lifecycle carbon footprint of BLOOM, a 176B parameter model — including training, deployment, and inference costs.
Google researchers quantified the carbon footprint of training large AI models — GPT-3 training emitted approximately 552 tonnes of CO2, equivalent to 123 gasoline cars driven for a year.