Exposure to body-positive imagery and humorous content parodying negative body stereotypes on social media inspires women to take action against unrealistic beauty standards, according to a new study from the University of Surrey.  

Social media use has long been part of the everyday lives of most children and adolescents. Many of them exhibit risky, and in some cases even addictive, behaviour. While social media use can certainly have positive effects for young people, intensive use can negatively impact mental, emotional, and social well-being, leading to symptoms such as depression and anxiety, impaired attention, and sleep problems. In a discussion paper published by the German National Academy of Sciences Leopoldina, the researchers involved therefore recommend applying the precautionary principle. In the paper “Social Media and the Mental Health of Children and Adolescents”, they give policy recommendations to protect children and adolescents from the negative effects of social media, for example by setting a minimum age for access or by restricting certain functions. The paper was published on 13 August 2025 and an English translation is now available.

Patients in specialised psychiatric care face a variety of social risks that are intertwined, a new study from the University of Eastern Finland shows. Social risks tend to accumulate among the most vulnerable segments of the population. They include for example financial difficulties, housing challenges and violence.

SUMMARY: ChatGPT appeared to “think” for itself and behave like a student – improvising ideas, making mistakes, and even resisting guidance – in a University of Cambridge study that revived a 2,400-year-old maths puzzle first recorded by Plato. The researchers asked the chatbot to tackle the classic “doubling the square” problem, which lays claim to being the oldest documented maths education challenge in history. Despite knowing about Plato’s work from its training data, ChatGPT improvised its own solution to the problem, sometimes making mistakes along the way. The study suggests that while this highlights that Generative AI has limits as a maths education resource, it may also have something resembling a  “zone of proximal development”: the potential to ‘learn’ and solve new – perhaps even unsolved – problems, with careful prompting and guidance.