Unbiased AI-powered news
Artificial intelligence systems have demonstrated stronger performance on mathematical tasks in recent tests. Researchers report that these advances still require human mathematicians to verify results and apply findings.
New ScientistArtificial intelligence tools have recorded measurable progress on advanced mathematical problems in recent months. Reports indicate that systems developed by technology companies can now address tasks previously viewed as requiring expert human skill.
Mathematicians have noted rapid improvement in AI capabilities within their field. One reported case involved an AI system disproving a mathematical statement that had stood for 80 years. The result gained acceptance only after review by human experts.
Human researchers have begun incorporating AI-generated suggestions into ongoing work on additional problems. These uses show that current AI output requires interpretation and validation steps performed by people trained in the discipline.
The pattern observed in mathematics suggests that broader scientific progress may follow a similar path. Human specialists select which AI outputs to pursue and integrate those outputs into larger research programs. Historical examples illustrate that major theoretical advances required teams of researchers to extend initial findings.
Without subsequent application by other scientists, even foundational work produced limited immediate effects on the wider field.
en.globes.co.ilAmazon CEO Jeff Bezos told attendees at VivaTech that he disagrees with concerns AI will make humans redundant.
EuronewsPrime Minister Sébastien Lecornu said the domestic intelligence service will switch providers to reduce reliance on foreign technology. The change follows a 2025 contract renewal and will take several years.
The IndependentNvidia's chief executive said society must adapt to artificial intelligence by creating new norms and safety standards. He compared the shift to changes that followed the arrival of automobiles.