Bb5a31f81ClaimAs of October 2026, the median author count for academic mathematics papers is two authorsAgreeDisagree
Bb5a31f81ClaimUntil the twentieth century, the majority of published mathematical research papers were single-authored rather than collaborative effortsAgreeDisagree
Bb5a31f81ClaimThomas Hales submitted his computational proof of the Kepler conjecture to the Annals of Mathematics in 1998AgreeDisagree
Bb5a31f81ClaimEthan Mollick co-authored a research paper titled 'Artificial Intelligence and Its Impact on Creativity and Diversity.'AgreeDisagree
Bb5a31f81ClaimA University of Chicago analysis of 41 million scientific papers found that AI use narrowed the range of scientific topics studied by almost 5%AgreeDisagree
Bb5a31f81ClaimA University of Chicago analysis of 41 million scientific papers found that AI use reduced researcher engagement with other scientists' work by 22%AgreeDisagree
Bb5a31f81ClaimCarl Friston, Glenn Fox, Michael Menino, and Steven Kotler published a paper titled The Body Doesn't Keep the Score in Frontiers in NeuroscienceAgreeDisagree
Bb5a31f81ClaimOpenAI released Rosalind Workbench, an agent suite designed to connect GPT-Rosalind to scientific instruments and data analysis workflowsAgreeDisagree
Bb5a31f81ClaimCitrini Research deployed 300 Claude subagents to evaluate 318,000 candidate proteins, selecting 211 candidates for laboratory testingAgreeDisagree
Bb5a31f81ClaimCitrini Research used Anthropic's Claude to design protein binders intended to neutralize scorpion neurotoxinsAgreeDisagree
Bb5a31f81ClaimIndependent laboratory testing confirmed that 17 AI-designed protein binders from Citrini Research bound to four distinct scorpion neurotoxinsAgreeDisagree
Bb5a31f81ClaimCitrini Research completed its computational protein design experiment for scorpion antivenom for less than several hundred thousand dollarsAgreeDisagree
Bb5a31f81ClaimThe AI4AD (Artificial Intelligence for Alzheimer's Disease) initiative utilizes machine learning across neuroimaging and multi-omic datasets to identify distinct pathological subtypes of dementiaAgreeDisagree
Bb5a31f81ClaimA Stanford study found that while humans rated AI-generated research ideas highly, human-generated ideas ultimately performed better in practiceAgreeDisagree
Bb5a31f81ClaimSpecializing in a narrow field is often the most effective way for postdoctoral researchers to secure publications and career advancementAgreeDisagree
Bb5a31f81ClaimAI is expected to accelerate progress in fields like life sciences by interfacing directly with experimentsAgreeDisagree
Bb5a31f81ClaimThe study on AI model costs was conducted by researchers from Stanford, Carnegie Mellon, UC Berkeley, and MicrosoftAgreeDisagree
Bb5a31f81ClaimAI models can now solve PhD-level physics problems and provide novel insights beyond the original researchAgreeDisagree
Bb5a31f81ClaimAI could accelerate scientific research to the point where a PhD-level project could be completed in weeksAgreeDisagree
Bb5a31f81ClaimAn iterative process involving AI, simulations, and physical experiments is necessary to advance scientific knowledgeAgreeDisagree