A new study reveals that civil servants, like citizens, do not uniformly recognize and oppose political attacks on liberal democracy. How they view such attacks and respond depends heavily on their social and political affiliation. Civil servants who recognize politicians’ actions as undermining democratic norms are more likely to withdraw—by resigning or reducing engagement—while others, aligned with the governing politicians, may see no problem, stay, and willingly cooperate. As a result, over time—especially under prolonged political pressure—the civil service may become less politically diverse, not only due to top-down politicization, but through patterns of voluntary exit and disengagement. This increases the incentive of subsequent governments to replace career civil servants with political appointees.

In India, the Slam Out Loud program connects teachers and artists in classrooms for storytelling, theater, and visual arts that bolster children’s socio-emotional learning. In the United States, Carnegie Hall partners with more than 115 orchestras across the country to teach children to sing and play instruments, culminating in an orchestral performance. Nonprofits like these are providing opportunities across the world in the face of reduced funding and support for arts education in public school systems.

The language capabilities of today’s artificial intelligence systems are astonishing. We can now engage in natural conversations with systems like ChatGPT, Gemini, and many others, with a fluency nearly comparable to that of a human being. Yet we still know very little about the internal processes in these networks that lead to such remarkable results.

A new study published in the Journal of Statistical Mechanics: Theory and Experiment (JSTAT) reveals a piece of this mystery. It shows that when small amounts of data are used for training, neural networks initially rely on the position of words in a sentence. However, as the system is exposed to enough data, it transitions to a new strategy based on the meaning of the words. The study finds that this transition occurs abruptly, once a critical data threshold is crossed — much like a phase transition in physical systems. The findings offer valuable insights for understanding the workings of these models.