How can society police the global spread of online far-right extremism while still protecting free speech' That's a question policymakers and watchdog organizations confronted as early as the 1980s and '90s ' and it hasn't gone away. Decades before artificial intelligence, Telegram and white nationalist Nick Fuentes' livestreams, far-right extremists embraced the early days of home computing and the internet. These new technologies offered them a bastion of free speech and a global platform. They could share propaganda, spew hatred, incite violence and gain international followers like never before. Before the digital era, far-right extremists radicalized each other primarily using print propaganda. They wrote their own newsletters and reprinted far-right tracts such as Adolf Hitler's 'Mein Kampf' and American neo-Nazi William Pierce's 'The Turner Diaries,' a dystopian work of fiction describing a race war. Then, they mailed this propaganda to supporters at home and abroad. I'm a historian who studies neo-Nazis and far-right extremism. As my research shows, most of the neo-Nazi propaganda confiscated in Germany from the 1970s through the 1990s came from the United States. American neo-Nazis exploited their free speech under the First Amendment to bypass German censorship laws. German neo-Nazis then picked up this print propaganda and distributed it throughout the country....

Scientists are increasingly using artificial intelligence (AI) to do their work. Many say the tools are saving them time and money, but others have seen the negative effects that such tools can have on research. In a survey of more than 2,400 researchers released in October by the publishing company Wiley, 62% of respondents said they used AI for tasks related to research or publication ' up from 45% in 2024, when there were 1,043 respondents. Early-career scientists and researchers in physical sciences were the most likely to use AI tools in their work, and were more likely to be early adopters of AI than were later-career researchers or those working in humanities, mathematics or statistics. Researchers are using AI tools to help with writing, editing and translating. They are also using them to detect errors or bias in their writing, and to summarize large volumes of studies. In a sample of 2,059 respondents, 85% said AI helped with efficiency, 77% that it helped to increase the quantity of work completed, and 73% that it improved the quality of their work....

Reducing the visibility of polarizing content in social media feeds can measurably lower partisan animosity. To come up with this finding, my colleagues and I developed a method that let us alter the ranking of people's feeds, previously something only the social media companies could do. Reranking social media feeds to reduce exposure to posts expressing anti-democratic attitudes and partisan animosity affected people's emotions and their views of people with opposing political views. I'm a computer scientist who studies social computing, artificial intelligence and the web. Because only social media platforms can modify their algorithms, we developed and released an open-source web tool that allowed us to rerank the feeds of consenting participants on X, formerly Twitter, in real time. Drawing on social science theory, we used a large language model to identify posts likely to polarize people, such as those advocating political violence or calling for the imprisonment of members of the opposing party. These posts were not removed; they were simply ranked lower, requiring users to scroll further to see them. This reduced the number of those posts users saw....

Chinese artificial intelligence company DeepSeek has released a mathematical reasoning model that can identify and correct its own errors. The model beat the best human score in one of the world's most prestigious undergraduate maths competitions. The model, DeepSeekMath-V2, scored 118 out of 120 points on questions from the 2024 William Lowell Putnam Mathematical Competition, beating the top human score of 90. The model also performed at the level of gold-medal winners in the International Mathematical Olympiad (IMO) 2025 and the 2024 China Mathematical Olympiad. The results are described in a preprint1 posted on arXiv on 27 November. In February, AlphaGeometry 2, an AI problem solver created by Google DeepMind in London, also achieved a gold-level performance in the IMO. The feat was repeated in July by Gemini's Deep Think, which is owned by DeepMind. Early approaches to training large language models for mathematical reasoning focused on the accuracy of final answers, the preprint authors write. But a correct answer does not guarantee correct reasoning. At times, a correct final answer might just be a result of a fortunate error. Moreover, an exclusive focus on the end result is not useful in proving mathematical laws or formulae, when the logical reasoning is more important than the final answer....