Philosophical Transactions of the Royal Society B

Abstract Objective Ambient AI scribes evaluated using frequency-based metrics such as word error rate (WER), which do not represent clinical consequence. We tested whether variation in these metrics tracks consequential transcription errors. Methods We constructed a multilingual corpus from five clinical dictation scripts spanning a complexity gradient, translated into 99 languages, rendered to synthetic speech under three acoustic conditions, and transcribed by a production ambient scribe. Six frequency metrics were computed. Three independent large language model raters from external providers assessed clinically meaningful error patterns in context using a Severity x Likelihood framework informed by UK digital clinical-safety-risk-management principles. Results Across 59,819 genuine transcription-error occurrences, 58,329 (97.5%) were LOW risk and 251 (0.42%) CRITICAL or HIGH. None of six frequency metrics showed a statistically detectable association with serious clinical risk across languages; correlations were small (absolute Spearman {rho}<0.16). A Severity x Likelihood sum remained strongly correlated with WER ({rho}=0.80), showing that the aggregate remained dominated by benign errors. At complexity level 3, low-resource languages had worse WER than high-resource languages ({beta}=+0.078, 95% CI +0.045 to +0.111; p<0.0001), without a detectable difference in CRITICAL/HIGH risk (OR 1.21, 95% CI 0.43 to 3.43; p=0.72). Consultation complexity was the principal predictor of serious risk (OR 3.06 per level, p<0.0001). Conclusion Across this controlled multilingual corpus, aggregate transcription-frequency metrics did not reliably track the sparse severe tail of clinically consequential errors. WER remains appropriate for transcription quality, but these data do not support its use alone as a proxy for clinical safety. Context-aware assessment of error consequence provides complementary information that frequency measures can dilute.

Practitioner-derived taxonomies of conversational artificial intelligence (AI) workflow errors show low inter-rater agreement among human coders, leaving open whether the instrument is ill-specified or the judgments are inherently difficult. We delivered a locked eight-category workflow-error taxonomy verbatim, under standardized conditions, to three frontier large language model comparators from distinct developer lineages, which coded a documented 45-incident error corpus. On the same 16 incidents coded by three human reviewer-authors, comparator category agreement was substantial (Fleiss {kappa}=0.625) against slight human agreement ({kappa}=0.155); across the full corpus it was stable ({kappa}=0.632), and a 10-category refinement did not reduce it ({kappa}=0.690). Severity and a claimed-verification flag remained only fair in both arms and on both samples. Substantial cross-system consistency provides a legibility signal consistent with recoverable category distinctions, but cannot separate instrument clarity from shared model priors, and is not a validation of any coding.

The Mutational Hazard Hypothesis (MHH) predicts that reduced effective population size weakens purifying selection and promotes transposable element (TE) accumulation. Because effective population size is difficult to estimate across broad taxonomic scales, body mass, generation time, and dN/dS are often used as proxies. We analyzed TE landscapes across 167 sauropsid genomes to test whether these proxies predict genomic TE proportion consistently across phylogenetic scales. All three showed strong scale dependence. Body mass was positively associated with TE proportion across clades, but this relationship broke down within clades: only snakes retained a significant negative Pearson correlation after multiple-testing correction, whereas the corresponding phylogenetically corrected slopes were not significant. Generation time showed a strong pooled association that disappeared within every major clade in both Pearson and PGLS analyses. The pooled dN/dS-TE relationship also disappeared after accounting for body mass and generation time in a structural equation model, with a robust within-clade association retained only in turtles. Lineage-specific TE dynamics, including LTR expansion in sea snakes, were not captured by these proxies. These results show that commonly used MHH proxies mainly reflect clade-level structure rather than consistent within-lineage mechanisms.

Eponymous etymologies (i.e., organisms named after people) are prevalent in global taxonomy worldwide, across scientific and common names. The persistent usage of eponyms has been intensely debated in recent times, across various sociopolitical, applied and philosophical perspectives. However, a rigorous statistical analysis of eponyms and their enduring colonial legacies remains hitherto unexplored. Here, we use a global and spatially-explicit network analysis to examine 2416 contemporary bird eponyms (21% of all bird species on earth) and their relationships with historical European colonialism. Additionally, we develop and utilise a Colonial Exposure Index (CEI) that summarizes the geopolitical influence of various colonial empires on modern-day countries. Our results show that an overwhelming majority of bird eponyms worldwide are named after people belonging to countries where the bird species does not occur. A majority of these eponyms honor scientists, aristocrats and army officers from erstwhile colonial empires, with eponymous species descriptions peaking during colonial expansion. Importantly, our analysis highlights a positive relationship between the Colonial Exposure Index (a metric factoring the duration and area of a country colonized by a specific empire) and the number of eponymous bird species honoring persons from that specific empire. Our temporal analysis, however, shows that the use of eponyms has recently shifted towards honoring native persons and distinguished conservation champions. Overall, this study presents an important statistical examination of eponyms and their colonial legacies, thereby providing valuable context for ongoing debates on the reform or retention of eponyms worldwide.

Integrating auditory and visual cues is a hallmark of human speech perception, yet adults from East Asian backgrounds show less reliance on visual speech than their Western counterparts. The origins of this cultural difference, however, remain unknown. To investigate whether this divergence is established early in infancy, we examined audiovisual integration in 6- to 12- month-old White Canadian (n=111) and Chinese (n=115) infants using a novel paradigm measuring their perception of the McGurk effect. Across four experiments, we found a clear developmental divergence: Canadian infants showed a stable McGurk effect from 6 months onward, whereas Chinese infants showed a more protracted developmental trajectory, a cultural pattern that was further highlighted when their integration was challenged by other-race faces. These findings provide the first direct evidence that cultural differences in multisensory speech perception are established within the first year of life, suggesting that the brains strategy for binding sight and sound is shaped by early experience, with broad implications for theories of language acquisition and developmental science.

Island biogeography is the theoretical and empirically validated expectation that an islands biodiversity is ultimately limited by its size and isolation, with larger and less isolated (from mainland communities) islands supporting greater diversity. Island biogeography is generally applied to simple metrics of diversity such as species richness (number of different species); however, fewer studies have used traits to measure biodiversity. An organisms traits -- e.g., body mass, age at sexual maturity, trophic level -- can be used to measure biodiversity and understand how ecological communities function. We quantified species richness for birds, mammals, reptiles, and amphibians, and functional diversity for birds and mammals (for which sufficient trait data were available), and tested whether this diversity can be predicted using the theory of island biogeography. We identified 9,103 Australian islands, of which 1,661 had at least one (native and/or non-native) non-marine species present according to the Atlas of Living Australia. As expected, tetrapod species richness (S) increased with island area (A) (z = 0.299 {+/-} 0.012) following a typical power-law relationship (i.e., S = cAz, where z = 0.2-0.4), but was not predicted by distance from mainland -- consistent with the pattern observed on other recently (< 10,000 years) isolated continental islands. We found that trait richness increased at the same rate with island area as species richness for mammals, but for birds, trait richness increased more slowly than species richness. Trait turnover increased modestly with inter-island distance, whereas trait nestedness was unrelated to distance. Trait richness increased strongly with species richness in both birds and mammals, and island area and isolation explained no additional variation in functional richness after accounting for species richness. These results indicate that island geography influences functional diversity primarily through species accumulation, rather than through direct effects on occupied trait space. Overall, the trait space of smaller islands tended to be nested within that of larger, nearby islands; the main differences among similar-sized islands are due to turnover (change in species/trait combinations among assemblages), and the effects are more pronounced in mammals compared to birds. We also found evidence for an asymptotic relationship between trait and species turnover in both birds and mammals, suggesting close coupling between taxonomic and functional turnover, with some saturation of trait turnover at high species turnover. Large islands that are simultaneously more isolated might offer conservation advantages by reducing the influence of threatening processes on the mainland if distance limits access of people and invasive species.

Rapid evolution allows populations to persist in environments where they would otherwise go extinct. This phenomenon, known as evolutionary rescue, is typically studied in the framework of biological evolution, yet adaptive traits can also arise and spread through cultural evolution. The present study developed a stochastic eco-evolutionary model to compare rescue probabilities through biological and cultural evolution. Transmission bias governed the rescue probability under cultural evolution by setting how readily a rare adaptive trait was copied. Conformity bias suppressed population persistence because a rare trait was the least likely to be copied. Content bias toward the adaptive trait enabled evolutionary rescue when social learning was rapid, but it typically yielded a lower rescue probability than biological evolution. Only anticonformity bias, together with a high social learning rate, exceeded the rescue probability of biological evolution by enabling the adaptive trait to be established more rapidly. These results demonstrate that transmission bias alters the demographic consequences of cultural evolution and highlight the importance of transmission processes in evolutionary rescue theory. Understanding how adaptive behaviours are socially transmitted may also improve predictions of animal population persistence and inform conservation efforts in rapidly changing environments.

Anthropogenic climate change is threatening global biodiversity, with sea turtles particularly vulnerable as offspring sex and developmental success are strongly influenced by incubation temperature. Behavioral plasticity, including the seasonal distribution of reproductive output, may provide short-term mechanisms for mitigating these impacts. Here, we investigated season-wide hatchling sex ratios and emergence success in a small population of green turtles (Chelonia mydas), tracking individual females across their nesting seasons. Sex ratios varied markedly through the nesting season, with later nests producing a greater proportion of male hatchlings. Moreover, sex ratios were relatively consistent among nests laid by individual females. Overall, females producing more nests over a season also produced more male offspring, suggesting that both nesting phenology and reproductive output influence individual contributions to future population demographics. Mechanistic models indicate that hatchling sex ratios have trended towards female-biased ratios (>80% female) over the past 50 years, and are projected to approach complete feminization by 2100 under continued warming. Although emergence success currently remains high (>85%), it is predicted to decline sharply after mid-century, with viable hatchling production falling to [~]30% by the end of the century. Models further show that maintaining contemporary sex ratios and emergence success will require unrealistically large delays in nesting phenology, and that even extreme shifts in phenology become ineffective by 2100. Together, these findings demonstrate that individual females can increase male hatchling production by nesting later and producing more nests, but behavioral plasticity alone is unlikely to offset the accelerating impacts of climate change on this population.

The sex ratio at birth (SRB), or secondary sex ratio, is stable, with more boys born alive, but shifts after major events. Acute stress may reduce SRB through disproportionate male fetal loss, while sexual intercourse patterns around conception are associated with offspring sex. We examined whether Brazil's April 2021 Gamma-wave mortality peak was followed by an SRB decline within the previously reported stress-sensitive 3-5 month window. We analysed >81 million live births (1997-2024) from Brazil's Live Birth Information System (SINASC) using pre-pandemic autoregressive forecasts with 95% prediction intervals (PIs). In July 2021, SRB was 50.951% versus 51.185% expected, below the 95% PI (50.968-51.402%), indicating fewer boys. All five macroregions were below expectation, largest in the North, including Amazonas, where Gamma was first identified. Monte Carlo simulation showed that, under usual SRB patterns, a decline with this timing and consistency across Brazil would occur about once in 200 simulations (probability 0.00467), making July 2021 unusual. In February 2023, SRB was 51.441% versus 51.220% expected, above the 95% PI (51.003-51.438%), indicating more boys, nine months after Brazil ended its COVID-19 emergency, suggesting increased sexual intercourse after pandemic relaxation. In April 2023, SRB was 50.956% versus 51.221% expected, below the 95% PI (51.004-51.439%), indicating fewer boys, four months after Pele's death. A comparable UK decline followed Princess Diana's death 4-5 months later. Gamma was prespecified; later observations are hypothesis-generating. These findings highlight SRB's population-level utility and warrant investigation of stress-related and behavioural pathways; ecological design precludes causality.

Artificial intelligence (AI) assistants can improve individual performance but may weaken unaided performance and limit interest diversity, raising questions about their long-term effects on collective intelligence (CI). Here, we consider an evolutionary model in which individuals predict, consult AI, and imitate better-performing peers. Classic reward structures that promote CI without AI can fail under Chatbot AI, whose predictions depend on users' interests, by reducing interest diversity or propagating bias. Directly rewarding or penalizing AI use does not reliably mitigate these effects. Instead, a balanced incentive that preserves minority expertise and rewards collective contributions can promote CI, but an additional incentive is needed to restore unaided collective accuracy. Our framework offers an explanation for emerging evidence of AI's societal impact and clarifies its long-term implications.

Consensus is rarely a one-off achievement. In a changing world, groups must repeatedly update judgments they have already agreed upon, making consensus updating a common feature of real-world collective decisions. However, previous research on collective decision-making has focused mainly on how groups reach a single consensus, leaving the updating of established consensus largely unexamined. This raises two unanswered questions: does consensus updating involve biases analogous to sunk cost, and are such biases related to the cost of the previous consensus? Building a Fake Consensus paradigm, we manipulated consensus-formation cost and examined how dyads updated a subjectively established consensus after new information arrived. An independent individual experiment provided information-matched updating baselines, allowing us to determine whether dyads updated less or more than individuals responding to the same evidence. Contrary to sunk-cost predictions, low-cost consensus led dyads to update less than the individual baseline, whereas medium- and high-cost consensus led them to update more. The pathway through perceived consensus sharedness was significant, and the association between sharedness and updating was stronger at higher left temporoparietal junction inter-brain synchrony. These findings indicate that the history of consensus formation shapes subsequent updating by altering the psychological meaning of shared endorsement, while interpersonal alignment determines how this shared experience translates into coordinated behavior. To our knowledge, this is the first experimentally controlled study of consensus updating in teams to isolate formation history from actual disagreement, providing a new paradigm for studying how collective judgments change after agreement.

Multi-agent prompting of large language models has produced contradictory diagnostic results, and it remains unclear whether any benefit comes from specialist personas or from perspective independence. We compared a single direct call, five personas in one context, and the same five roles as isolated agents integrated by a moderator, with five repeat runs per case, on 87 CPC cases, 406 MedCaseReasoning cases, and 364 emergency department encounters, under an LLM judge validated against clinicians. On the external benchmark the team beat the single call on both pre-specified recall endpoints (top-3 +3.0 points, p = 0.0079; top-5 +3.9, p = 3.8 x 10^-5;); a factorial attributes the gain to independent generation plus moderated synthesis, not the specialist roles. On real emergency presentations the benefit reversed (top-1 40.1% versus 34.3%, p < 0.0001), carried by the specialist role lists and surviving added objective results. Deployment should key on the question and the input at hand.

Primates maintain cooperative relationships largely through grooming reciprocity, yet studies disagree on the timescales over which grooming is returned. Part of this disagreement stems from statistical issues: inconsistent definitions of a reciprocation window, combined with the exclusion of unreciprocated grooming, leave the debate unresolved. We still do not know if primates exchange grooming within the short windows predicted by biological market theory or over much longer timescales. Whilst survival analysis can identify reciprocation windows objectively by retaining unreciprocated grooming as censored data, previous studies have applied it inconsistently. This inconsistency leaves the true time frames over which primates reciprocate grooming somewhat blurred. Here, we use a statistical approach that addresses these problems to examine grooming reciprocity in crested macaques (Macaca nigra), a socially tolerant macaque species. We applied survival analysis to identify each group's short-term reciprocation window without imposing arbitrary cut offs, and Bayesian survival models to test whether social bond strength and eigenvector centrality predicted reciprocation latency. Short-term reciprocation occurred rarely (29% of events in one group and 5% in the other); most reciprocation occurred over longer timescales than previous research commonly assumes. Stronger bonds predicted faster reciprocation; this finding suggests that crested macaques express social investment in valued partners through how promptly they return grooming, meaning that the timing of exchange carries information about relationship value. Eigenvector centrality showed no credible effect, though our small sample size and dense groups prevent us from interpreting this outcome more robustly. Substantial variation between dyads remained unexplained, pointing to partnerspecific properties we have yet to identify. These findings caution against assuming that reciprocity operates over short windows. They show that social predictors, such as bond strength, remain essential to understanding when primates return grooming. This work provides a baseline for testing whether grooming is exchanged for other social commodities, and for comparing reciprocation timescales across macaque species that differ in social tolerance.

Classic socio-ecological theory predicts that males and females experience different sources and mechanisms of social competition. Whether these differences translate into sex-specific structural properties of dominance hierarchies remains unclear. Here, we compiled 156 dominance interaction matrices from 80 published studies and extracted three commonly used metrics - hierarchy steepness, linearity and the directional consistency index - to investigate the structural characteristics of male and female dominance hierarchies across primates. All three metrics were strongly affected by methodological and demographic variables. Steepness increased with the number of recorded interactions and group size, linearity decreased as matrices became sparser, and directional consistency declined with increasing numbers of interactions. Steepness covaried positively with both linearity and directional consistency, indicating that groups with steeper hierarchies also exhibited more linear and more directionally consistent relationships. We found no sex differences in steepness, linearity or directional consistency. These results suggest that current metrics primarily reflect variation in the sampling effort and the rate of interaction of the recorded behaviour and appear therefore not to capture potential sex differences in the forms of competition. Our findings highlight the need for alternative measures of power asymmetries that are less confounded by sampling effort and demographic variation to better understand how competition and conflict are structured across primate societies.

Compositional syntax - the ability to combine meaningful units into structured sequences whose overall meaning depends on both the meanings of the parts and their combination - is a hallmark of human language. Recent work suggests that chimpanzees may exhibit similar capacities, raising the possibility that compositionality predates the emergence of language in humans. However, a key challenge has been to exclude simpler explanations in which call combinations are processed as independent utterances rather than integrated structures. Here, we address this issue using a source-manipulation paradigm in playback experiments with captive chimpanzees (n = 6) at the Sigean African Reserve, France. We presented sequences combining alarm and recruitment calls either from a single source (one loudspeaker) or from two spatially distinct sources (two loudspeakers). If sequences are interpreted as independent calls, responses should not differ between conditions. Instead, chimpanzees responded significantly faster when sequences originated from a single source. Bayesian survival analyses confirmed this effect. These results support a non-trivial form of compositionality, whereby the meaning of call combinations depends on their integration into a unified sequence produced by a single individual. Our findings provide experimental evidence that key features of linguistic compositionality may have evolutionary roots in the primate lineage.

Beta diversity is a fundamental component of regional biodiversity that accumulates from compositional differences between local species assemblages. Although beta diversity is increasingly applied across spatial scales, its scale dependence remains incompletely synthesized, particularly concerning variation in spatial scale used to define alpha diversity (grain size). Using tree community data from Barro Colorado Island, we find that both additive and multiplicative beta diversity are consistently higher for unequal than equal grain divisions across different spatial extents and numbers of grains. These results can be explained by Jensen's inequality based on the concave-down relationship of alpha diversity with area that in turn increases beta diversity through the scaling relationship between alpha and gamma diversity. Combined with our additional analysis showing that this prediction is robust to more general models of species richness, including species--area and species--sampling-effort curves, these results reveal a previously unrecognized, predictable range of beta diversity variation.

Historical biogeography seeks to understand the drivers of species distributions over space and time. One question of interest is how, out of many possible ways, does geography drive speciation. Vicariance, where geographic barriers arise splitting populations and limiting gene flow, can lead to allopatric speciation. Founder events, where a small number of individuals disperse over a barrier, can similarly lead to allopatric speciation if the individuals remain isolated. Both of these scenarios can lead to identical ranges of and relationships between modern species. Classic biogeographic approaches often focus on the history of populations of one or few species on shallow time scales or multiple species on deep time scales. We argue that focusing exclusively on either end of this time spectrum misses a venue for investigating the biogeographic drivers of speciation, at least for certain speciation events. With simple coalescent simulations with multiple species, we show as a proof of concept that gene tree distributions vary predictably between vicariance and founder event speciation. The existence of predictable patterns warrants the development of new approaches that capitalize the gene trees to distinguish vicariance and dispersal as drivers of speciation.

Human interlocutors have strong expectations that guide when, how, and what is communicated, but the evolutionary origins of such normative expectations remain poorly understood. Vocal accommodation in marmosets, where individuals modify their calls to match partners, offers a potential model. Within an interactive closed-loop playback paradigm, marmosets experienced partner calls that either converged toward or diverged away from their own call structure. Marmosets responded affiliatively to convergence, whereas diverging calls elicited protest vocalizations. Importantly, the protests stopped when the calls converged again. Effects were partner-specific, indicating that marmosets expect bond partners, but not strangers, to converge during a conversation. Furthermore, individuals modified their calls to restore similarity to dissimilar partner playbacks. These findings identify marmoset vocal accommodation as a socially regulated behavior with normative properties.

The consequences of maternal effects for offspring fitness can depend on the environmental conditions offspring encounter and the fitness traits considered. We used a transgenerational experiment in Daphnia magna to test how maternal exposure to warming and elevated salinity influenced offspring survival, reproduction and body size across matching, mismatching and multiple-stressor environments. Maternal exposure to salinity enhanced offspring survival under matching salinity, and benefits extended to mismatching environments. In contrast, maternal exposure to warming provided no survival benefit under matching warming and reduced survival in mismatching environments, while accelerating reproduction across contexts. Multiple-stressor conditions revealed pronounced maternal effects, while contrasting responses across survival, reproduction and body size showed that benefits to one fitness trait did not extend to others. Our findings demonstrate that environmental matching alone is insufficient to predict maternal effects, whose consequences depend on offspring environmental complexity and the fitness trait considered.

Although cities are increasingly recognized as ecological islands, a unified framework explaining their susceptibility to alien plant invasion remains lacking. Using the most recent and comprehensive global dataset of urban alien plants, we modelled alien richness, mimicking island biogeography theory (IBT). Across all models, neither city size nor geographic isolation independently explained alien richness. Instead, richness was consistently associated with their interaction, supporting the central IBT prediction. However, the strength of this interaction depends on how city size was quantified, with socio-economic dimensions exhibiting stronger positive interactions with geographic isolation than physical measures of city size. Introduction-hub identity further modified these relationships. North America was the only hub for which the interaction between city size and isolation was consistently weakened, indicating that donor regions of alien plants are not ecologically equivalent. Simulations of simultaneous increases in city size and isolation showed that larger, more connected cities generally accumulated more alien plants despite increasing geographic distance, but the magnitude and direction of these responses are hub dependent. Our findings inspire an extension of classical IBT to a mechanistic explanation for global variation in urban alien plant richness in this increasingly urbanized and globally connected world.