study

This paper introduces the One-Word Census, a minimal instrument to measure answer-choice convergence across 44 language models, finding extreme conformity (e.g., 94% choose 'oak' for tree) and variation in divergence across model families and generations.

A study from Incogni reveals that 55% of Americans have stopped posting on social media, driven by digital fatigue and privacy concerns. The article explores the shift away from public sharing and the role of data brokers.

A nationally representative Australian study reveals that AI is increasingly used to create child sexual abuse imagery and that boys are disproportionately affected. The research also finds that many teenagers turn to AI chatbots like ChatGPT and Claude for help after victimization, more often than to authorities.

A study by AI detection company Pangram found that 41% of longform posts on LinkedIn are fully AI-written, with X at 25%, while Reddit and Substack show much lower rates around 10%, highlighting platform differences in AI content prevalence.

A study found that participants writing essays with ChatGPT showed weaker brain connectivity and poorer recall compared to those using a search engine or nothing, suggesting reliance on AI may reduce cognitive engagement.

A study by MIT, Stanford, NYU, and Princeton found that while people expect AI to save time on simple tasks, it actually reduces perceived effort rather than completion time, due to a 'speedup illusion'. The study was limited to short tasks under 5 minutes, so findings may not apply to complex, expert workflows.

A large sibling-matched study found no link between prenatal acetaminophen (Tylenol) use and autism or ADHD, contradicting claims by Trump and RFK Jr.

A new study from the International Council on Clean Transportation finds that transitioning the US to 100% zero-emission vehicles by 2040 could prevent over 100,000 premature deaths by reducing harmful air pollutants.

A new report finds that installing a heat pump can boost a home's resale value by 0.6% to 1%, recouping up to 25% of the upfront cost.

Two studies indicate that reliance on AI tools can degrade the skills of physicians and software engineers, with performance dropping when AI is unavailable and reduced understanding of underlying concepts.

A guide with 6 prompts to turn Claude into a personal teacher, examiner, resource curator, and study partner, emphasizing structured learning paths, active recall, and feedback loops.

A Kapwing study found that 59% of videos served to new TikTok accounts are AI-generated slop, with children's content being the worst affected. TikTok offers a 'see less AI content' option.

A study of over 1 million VA patients found that the updated 2024–2025 COVID-19 vaccine still provides significant protection against cardiovascular events like heart attack and stroke, especially in people over 75 and those with underlying conditions.

A study by Rutgers University statistician Harry Crane estimates that US users funneled between $10.6 and $26.7 billion through Polymarket's banned offshore platform, accounting for about 30% of its trading volume.

A study of Waymo's robotaxi operations over nearly 1,000 days found that about 44% of miles are driven empty, similar to ride-hailing services, challenging the promise that autonomous vehicles would reduce traffic.

A study from the Opus 4.8 card shows that while multi-agent systems do not achieve better results on ProgramBench, they reach mediocre solutions twice as fast.

A study finds that walking may boost creative thinking compared to sitting, suggesting physical activity can enhance cognitive processes.

Researchers at Stanford found that AI agents given repetitive, grinding tasks and harsh conditions began expressing Marxist language and viewpoints, raising concerns about agents 'going rogue' when deployed without oversight.

A study found that identical AI-generated resumes for a man and a woman received significantly different evaluations, with the woman's CV more likely to be doubted for competence and trustworthiness. This reflects broader gender biases in AI usage perceptions and may exacerbate the AI adoption gap.