Experts have unveiled a new proposal for the governance of artificial intelligence. The new framework, published in the journal AI and Ethics, introduces the “Four As” – Action, Access, Authority and Allocation – and is the work of Dr Kirti Prakash and Dr Ian van der Linde of the Advanced Computing Research Centre at Anglia Ruskin University (ARU).

The “Four As” framework is intended to preserve competition and innovation in frontier AI development while establishing safeguards around capabilities and uses for AI that could cause severe or irreversible harm. The framework draws lessons from previous attempts to govern powerful technologies, including nuclear arms control, CRISPR governance around genome editing and the 1975 Asilomar conference on recombinant DNA technology. 

University of Florida's Sumit Jha, Ph.D., has developed the ATLAS system, which scans an LLM and senses when the model begins to drift from the task and redirects. 

CU Anschutz School of Medicine radiologist Harrison Bai, MD, MS, is part of a team that designed an AI tool to interpret right upper quadrant (RUQ) ultrasounds more quickly and accurately, saving doctors time and helping patients get speedier care.

Premature birth may do more than interrupt brain development – it may fundamentally alter how the cerebellum matures into early childhood, according to a new study. The findings reveal how the timing of birth shapes brain cell development and neural connections, offering new insight into the lasting neurodevelopmental effects of prematurity. The cerebellum undergoes its most rapid period of growth between 24 and 40 weeks of gestation, making premature birth a potential disruption to a critical phase of brain development. Although early birth is known to alter cerebellar growth and neurodevelopment, the underlying molecular and cellular mechanisms remain poorly understood. To address this gap, Georgios Sanidas and colleagues studied living and postmortem infants, some of whom were born prematurely, while others were born at term. They combined postmortem cerebellar tissue analysis with brain imaging and developmental assessments.

Using spatial gene-expression profiling and deep-learning-based tissue analysis, Sanidas et al. mapped cerebellar development across a broad range of gestational ages. The analysis revealed that premature birth reshapes the cerebellum’s developmental trajectory, with the timing of birth influencing how distinct brain cell populations mature even after infants reach the same postmenstrual age as those born at term. According to the authors, premature development involves disrupted coordination between Purkinje and granule cells, potentially impairing the refinement of neural connections despite the cerebellum retaining its overall structure. These findings suggest that prematurity does not simply interrupt normal development or expose the brain to injury; it alters the biological conditions under which maturation unfolds. Although the long-term persistence of these changes remains uncertain, the authors argue that the findings offer a framework for understanding how early birth influences later motor and cognitive outcomes, which may help guide the timing of future interventions. “Although the preterm human infant may follow two calendars, one for cerebellar physical growth and one for cerebellar cell and molecular maturation, the findings of Sanidas et al. show that neither captures the true complexity of brain health at the time, nor what it means for the child years later,” write Molly Kelly-Goss and Elizabeth Crouch in a related Perspective.

Bird songs share surprising similarities with human music and speech, according to a large-scale study comparing vocalizations across hundreds of bird species and dozens of human languages. The findings suggest that these similarities may arise partly from shared biological mechanisms, rather than vocal learning ability alone. Birdsong shares several important features with human speech and music, including complex sequences of sounds, coordinated vocal movements, and variations shaped by social context. These similarities suggest that bird vocalizations may offer valuable insights into the biological and evolutionary foundations of human communication. However, only by comparing a wide range of bird species can researchers reveal which patterns of vocal communication are common to birds and humans.

Logan James and colleagues compiled and analyzed a curated library of songs from 367 bird species, including both vocal learners (birds that learn their sounds) and nonlearners, and compared the sounds with human speech and song across 52 languages. James et al. found that several common features of human speech and music, including small pitch changes, regular syllable timing, and lengthened sounds at the ends of phrases, also occur across hundreds of bird species, regardless of whether they learn their vocalizations. These shared patterns suggest that fundamental aspects of vocal communication may arise from common motor and sound-production mechanisms rather than vocal learning alone. However, differences between birds and humans, as well as variation among bird species, indicate that factors such as body size, respiratory physiology, and vocal anatomy also shape how sounds are produced and organized. The findings further reveal that some bird groups more closely resemble humans in certain aspects of vocalization, potentially making them useful models for investigating the biological foundations of speech and music. “James et al. provide a basis for testing the mechanisms behind shared vocal patterns. Linguistic concepts can guide such tests but should not be used in place of experimental data,” writes Kazuo Okanoya in a related Perspective, which discusses some of the study’s limitations.