Coordinated, effective and efficient management of invasive alien species (IAS) relies on access to standardised information on past and ongoing control efforts. However, existing data are typically fragmented, difficult to locate, inconsistently formatted, or remain unpublished – particularly in cases where management was less successful. This limits the potential for past efforts to inform the design of future management strategies and to identify knowledge gaps that could be filled by applied research. To help address these challenges, for example by increasing reporting rates and improving the quality of recorded data, we co-developed a unified data framework capable of capturing salient information on IAS removal projects. The framework is applicable to all types of IAS, management objectives (e.g. from containment to eradication), removal techniques and monitored outcomes, for any location and spatiotemporal scale, and for any degree of success. We produced a first draft of the framework by reviewing existing standards, databases, and reporting systems. This was refined through a series of online and in-person workshops involving >70 participants with expertise in IAS management from >30 countries. Finally, we tested the framework by extracting data for case studies from a variety of sources. The framework is split into sections capturing: project details, target IAS, goals, actions taken, economic costs and benefits, outcomes, reflections on implementation, and linked resources. The framework is deliberately comprehensive, containing >90 questions. However, the burden of data entry is reduced by conditional logic (some questions only appear if relevant based on previous responses), acknowledging that some data (e.g. detailed quantitative outcomes) can be captured by other linked frameworks, and highlighting a small subset of essential questions to allow for very rapid data entry. Our framework also uses preset response categories, borrowed from existing frameworks or platforms for interoperability, wherever possible. Areas of ongoing and future work include: (a) using the framework to populate a database of IAS removal projects, (b) building an online platform based on the framework and database, allowing users to interact with existing data and submit their own, (c) creating artificial intelligence tools to support data entry, and (d) developing crosswalks with terminology in other frameworks and platforms to maximise interoperability.

Wise use of evidence to support efficient conservation action is key to tackling biodiversity loss with limited time and resources. Evidence syntheses provide key recommendations for conservation decision-makers by assessing and summarising evidence, but are not always easy to access, digest, and use. Recent advances in Large Language Models (LLMs) present both opportunities and risks in enabling faster and more intuitive systems to access evidence syntheses and databases. Such systems for natural language search and open-ended evidence-based responses are pipelines comprising many components. Most critical of these components are the LLM used and how evidence is retrieved from the database. We evaluate the performance of ten LLMs across six different database retrieval strategies against human experts in answering synthetic multiple-choice question exams on the effects of conservation interventions using the Conservation Evidence database. We found that LLM performance was comparable with human experts over 45 filtered questions, both in correctly answering them and retrieving the document used to generate them. Across 1867 unfiltered questions, LLM performance demonstrated a level of conservation-specific knowledge, but this varied across topic areas. A hybrid retrieval strategy that combines keywords and vector embeddings performed best by a substantial margin. We also tested against a state-of-the-art previous generation LLM which was outperformed by all ten current models - including smaller, cheaper models. Our findings suggest that, with careful domain-specific design, LLMs could potentially be powerful tools for enabling expert-level use of evidence syntheses and databases in different disciplines. However, general LLMs used 'out-of-the-box' are likely to perform poorly and misinform decision-makers. By establishing that LLMs exhibit comparable performance with human synthesis experts on providing restricted responses to queries of evidence syntheses and databases, future work can build on our approach to quantify LLM performance in providing open-ended responses.

This document summarises the topics discussed during a visit to Chippenham Fen, hosted by reserve managers Chris Hainsworth and Mike Taylor (Natural England). The site at Chippenham is described along with proposed management actions, with inputs and suggestions from other attendees highlighted with an asterisk (*). Evidence from literary sources is provided in green boxes. This is not a detailed synthesis or comprehensive review, but rather an attempt to combine knowledge from experienced, local land managers with evidence from the Conservation Evidence database (www.conservationevidence.com) and other sources.

To halt and reverse the trends of ecosystem loss and degradation under global change, nations globally are promoting ecosystem restoration. Restoration is particularly crucial to coastal wetlands (including tidal marshes, mangrove forests, and tidal flats), which are among the most important ecosystems on Earth but have been severely depleted and degraded. In this review, we explore the question of how to make restoration more effective for coastal wetlands in light of the often-overlooked dynamic nature of these transitional ecosystems between land and ocean. Currently, restoration efforts have focused on removing anthropogenic threats, habitat reconstruction, and planting foundation species, often with mixed success and high costs. The challenges largely lie in the abiotic and biotic dynamics of these transitional ecosystems, including (i) fluctuating environmental stresses, (ii) variable trophic and nontrophic species interactions, (iii) changing connectivity with adjacent land, sea, and freshwater systems, and (iv) accelerating climate change, including sea level rise, droughts, and storms. Future restoration should explicitly account for these abiotic and biotic dynamics from threat removal to habitat reconstruction, assisted succession, and post-restoration management. We highlight novel yet practical measures to enhance success. In the coming decades, bending the curve of coastal wetland loss and degradation globally also requires better understanding of the abiotic and biotic dynamics of these transitional ecosystems prone to change, using this understanding to develop innovative restoration approaches, and applying new approaches to upscale restoration in synergy with socioeconomic development. Critical to these efforts are collaborations among ecologists, policymakers, business investors, restoration practitioners, and the many millions of people dependent on coastal wetlands.

Scientific publications often benefit from diverse contributions that go uncredited owing to a lack of guidelines for recognizing non-author contributors. We propose ‘extended research credits’ — a standardized, tiered system (modelled after the attribution style of the film industry) to highlight hidden labour in research.

This is a response to a [letter](http://doi.org/10.1016/j.tree.2025.03.003) by Murray K. et al to our paper on "[:2024-ai-conhorizon]". See [:ai-should-unite-conservation] for further thoughts.

The Internet has grown from a humble set of protocols for end-to-end connectivity into a critical global system with no builtin “immune system”. In the next decade the Internet will likely grow to a trillion nodes and need protection from threats ranging from floods of fake generative data to AI-driven malware. Unfortunately, growing centralisation has lead to the breakdown of mutualism across the network, with surveillance capitalism now the dominant business model. We take lessons from from biological systems towards evolving a more resilient Internet that can integrate adaptation mechanisms into its fabric. We also contribute ideas for how the Internet might incorporate digital immune systems, including how software stacks might mutate to encourage more architectural diversity. We strongly advocate for the Internet to “re-decentralise” towards incentivising more mutualistic forms of communication.

Butterflies and moths face a range of anthropogenic threats with many of the best-studied populations in decline. In response, butterfly and moth conservation programmes are implementing a diverse set of actions, but to date no study has synthesised evidence for their effectiveness. We present an overview of the recently published Conservation Evidence synopsis of butterfly and moth conservation, describe patterns and biases in the available evidence, and compare these to similar synopses on other taxa. We find that most evidence covers butterfly conservation, focuses on community-level responses, originates from the UK and the USA, comes from studies using the least robust designs, and assesses actions addressing the threat posed by agriculture. Far less evidence is available for moth conservation, for individual species’ responses, originates from Africa, Asia, Oceania and South America, comes from studies using the most robust designs, or tests actions designed to mitigate the impacts of pollution or climate change. While the geographic and study design biases reflect those found in evidence for the conservation of other taxa, the focus on community-level responses is higher than in any of the other synopses we examined. We suggest this may leave Lepidoptera conservation vulnerable to missing important, species-specific responses. We call for testing of conservation actions to be built into conservation projects for butterflies and moths, to build a robust evidence base for conserving Lepidoptera in a changing world.

Using relevant scientific evidence is crucial to effectively conserve species and ecosystems worldwide. Currently, evidence that is available only in non-English languages is severely underutilized. We examined many underutilized non-English languages in the conservation evidence literature and factors that facilitate the use of non-English-language evidence based on citation patterns of articles testing the effectiveness of conservation actions published in English and 15 non-English languages. Multivariate models incorporated explanatory variables, such as lexical distance from English, availability of an English abstract, study design complexity, conservation status of studied species, and language of citing articles. Non-English-language articles received significantly fewer English citations (i.e., citations in English-language articles) than English-language articles. Hungarian, Polish, Korean, and Russian articles were particularly undercited in English. Despite fewer English citations, many non-English-language articles had high citation rates in their own languages, indicating their value in local conservation communities. Non-English-language articles with English abstracts received more English citations. The content of the article, such as a more robust study design or assessment of threatened species, was not significantly associated with the number of English citations received. Our findings highlight the importance of increasing the visibility and recognition of non-English-language articles, especially those in currently underutilized languages, for a more comprehensive understanding of global conservation challenges. Providing a translated English abstract has the potential to increase readership of an article by increasing the accessibility to those who can understand English.

Understanding the consequences of past conservation efforts is essential to inform the means of maintaining and restoring species. Data from the IUCN Red List for 67,217 animal species were reviewed and analyzed to determine (i) which conservation actions have been implemented for different species, (ii) which types of species have improved in status and (iii) which actions are likely to have driven the improvements. At least 51.8% (34,847) of assessed species have actions reported, mostly comprising protected areas (82.7%). Proportionately more actions were reported for tetrapods and warm-water reef-building corals, and fewer for fish, dragonflies and damselflies and crustaceans. Species at greater risk of extinction have a wider range of species-targeted actions reported compared with less threatened species, reflecting differences in documentation and conservation efforts. Six times more species have deteriorated than improved in status, as reflected in their IUCN Red List category. Almost all species that improved have conservation actions in place, and typically were previously at high risk of extinction, have smaller ranges and were less likely to be documented as threatened by hunting and habitat loss or degradation. Improvements in status were driven by a wide range of actions, especially reintroductions; for amphibians and birds, area management was also important. While conservation interventions have reduced the extinction risk of some of the most threatened species, in very few cases has full recovery been achieved. Scaling up the extent and intensity of conservation interventions, particularly landscape-scale actions that benefit broadly distributed species, is urgently needed to assist the recovery of biodiversity.

The global food system, essential for delivering nutritional security to a growing population, is highly vulnerable to diverse hazards. This study investigates the feasibility of leveraging an existing systematic database, specifically the Conservation Evidence database, for mitigating environmental hazards impacting the food system. By focusing on human–wildlife conflict as a case study, we explored the database’s potential to inform hazard mitigation strategies. Our analysis revealed significant geographical and taxonomic gaps, varied intervention strategies and differences in study designs across regions. We identified key challenges, such as the need for comprehensive tagging and filtering features, integration of non-academic data and broader stakeholder engagement. The findings underscore the complexity of adapting conservation databases for food system applications but highlight the potential benefits of a free-to-access, systematic, evidence-based approach focusing on food production hazard mitigation. Future work should focus on developing a dedicated food system hazard database, leveraging automation and machine learning to enhance data extraction and application efficacy, ultimately improving global food security and sustainability.

The Internet has grown from a humble set of protocols for end-to-end connectivity into a critical global system with no builtin "immune system". In the next decade the Internet will likely grow to a trillion nodes and need protection from threats ranging from floods of fake generative data to AI-driven malware. Unfortunately, growing centralisation has lead to the breakdown of mutualism across the network, with surveillance capitalism now the dominant business model. We take lessons from from biological systems towards evolving a more resilient Internet that can integrate adaptation mechanisms into its fabric. We also contribute ideas for how the Internet might incorporate digital immune systems, including how software stacks might mutate to encourage more architectural diversity. We strongly advocate for the Internet to "re-decentralise" towards incentivising more mutualistic forms of communication.

The multibillion dollar ornamental plant trade benefits economies worldwide, but shifting and rapidly expanding globalized supply chains have exacerbated complex environmental, sustainability, and biosecurity risks. We review the environmental and social risks of this international trade, complementing it with analyses of illegal trade seizures and plant contaminant interception data from the Netherlands and the United Kingdom. We show global increases in ornamental plant trade, with supply expansions in East Africa and South America, and highlight risks and impacts including biodiversity loss, aquifer depletion, pollution, undermined access and benefit sharing, and food security. Despite risk mitigation efforts, the interception data showed considerable volumes of contaminants in ornamental plant shipments, but taxonomic identification was not always possible, highlighting uncertainties in assessing biosecurity risks. With high-volume and fast-moving transit of ornamental plants around the world, it is essential that production standards are improved and that data on specific risks from trade are collected and shared to allow for mitigation.

The publication of ever-larger numbers of problematic papers, including fake ones generated by artificial intelligence, represents an existential crisis for the established way of doing evidence synthesis. But with a new approach, AI might also save the day.

High-redshift protoclusters consisting of dusty starbursts are thought to play an important role in galaxy evolution. Their dusty nature makes them bright in the far-infrared (FIR)/submm but difficult to find in optical/near-infrared (NIR) surveys. Radio observations are an excellent way to study these dusty starbursts, as dust is transparent in the radio and there is a tight correlation between the FIR and radio emission of a galaxy. Here, we present MeerKAT 1.28 GHz radio imaging of three Herschel candidate protoclusters, with a synthesized beam size of similar to 7.5 arcsec x 6.6 arcsec and a central thermal noise down to 4.35 mu Jy beam (-1). Our source counts are consistent with other radio counts with no evidence of overdensities. Around 95 per cent of the Herschel sources have 1.28 GHz IDs. Using the Herschel250 mu m primary beam size as the searching radius, we find 54.2 per cent Herschel sources have multiple 1.28 GHz IDs. Our average FIR-radio correlation coefficient q(250 mu m) is 2.33 +/- 0.26. Adding q(250 mu m) a new constraint, the probability of finding chance-aligned sources is reduced by a factor of similar to 6, but with the risk of discarding true identifications of radio-loud/quiet sources. With accurate MeerKAT positions, we cross-match our Herschel sources to optical/NIR data followed by photometric redshift estimations. By removing z < 1 sources, the density contrasts of two of the candidate protoclusters increase, suggestive of them being real protoclusters at z < 1 There is also potentially a 0.9 < z < 1.2 overdensity associated with one candidate protocluster. In summary, photometric redshifts from radio-optical cross-identifications have provided some tentative evidence of overdensities aligning with two of the candidate protoclusters.

Trade represents a significant threat to many wild species and is often clandestine and poorly monitored. Information on which species are most prevalent in trade and potentially threatened by it therefore remains fragmentary. We used 7 global data sets on birds in trade to identify species or groups of species at particular risk and assessed the extent to which they were congruent in terms of the species recorded in trade. We used the frequency with which species were recorded in the data sets as the basis for a trade prevalence score that was applied to all bird species globally. Literature searches and questionnaire surveys were used to develop a list of species known to be heavily traded to validate the trade prevalence score. The score was modeled to identify significant predictors of trade. Although the data sets sampled different parts of the broad trade spectrum, congruence among them was statistically strong in all comparisons. Furthermore, the frequency with which species were recorded within data sets was positively correlated with their occurrence across data sets, indicating that the trade prevalence score captured information on trade volume. The trade prevalence score discriminated well between species identified from semi-independent assessments as heavily or unsustainably traded and all other species. Globally, 45.1% of all bird species and 36.7% of globally threatened bird species had trade prevalence scores ≥1. Species listed in Appendices I or II of CITES, species with large geographical distributions, and nonpasserines tended to have high trade prevalence scores. Speciose orders with high mean trade prevalence scores included Falconiformes, Psittaciformes, Accipitriformes, Anseriformes, Bucerotiformes, and Strigiformes. Despite their low mean prevalence score, Passeriformes accounted for the highest overall number of traded species of any order but had low representation in CITES appendices. Geographical hotspots where large numbers of traded species co-occur differed among passerines (Southeast Asia and Eurasia) and nonpasserines (central South America, sub-Saharan Africa, and India). This first attempt to quantify and map the relative prevalence in trade of all bird species globally can be used to identify species and groups of species that may be at particular risk of harm from trade and can inform conservation and policy interventions to reduce its adverse impacts.

AbstractUnderstanding the consequences of past conservation efforts is essential to inform the means of maintaining and restoring species. Data from the IUCN Red List for 67,217 comprehensively assessed animal species were reviewed and analysed to determine (i) which conservation actions have been implemented for different species, (ii) which types of species have improved in state and (iii) which actions are likely to have driven the improvements. At least 51.8% (34,847) of assessed species have actions reported, mostly comprising protected areas (82.7%), with more actions reported for both terrestrial tetrapods and warm-water reef-building corals and fewer for fish, dragonflies and damselflies and crustaceans. Species at greater risk of extinction have a wider range of species-targeted actions reported compared to less threatened species, reflecting differences in documentation and conservation efforts. Six times more species have deteriorated rather than improved in their Red List category. Almost all species that improved have conservation actions in place; species that improved in state typically were historically at high risk of extinction, have smaller ranges and lacked a range of reported threats, particularly hunting and habitat loss or degradation. All types of conservation action were associated with improvements in state, especially reintroductions and invasive species control, alongside, for amphibians and birds, area management. This suggests a range of conservation interventions have successfully conserved some species at greatest risk but have rarely recovered populations to resilient levels. Scaling up the extent and intensity of conservation interventions, particularly landscape-scale actions that benefit broadly distributed species, is urgently needed to assist the recovery of biodiversity.

This chapter reproduces the second horizon scan for bioengineering conducted by the Centre for the Study of Existential Risk in 2020, one of the most significant pieces of horizon scanning that has been undertaken in the field to-date. Identifying the top 20 emergent issues, the authors group them according to a likely timeline for their realisation, and discuss each throughout the chapter. This allows for the most notable issues that may impact the planet and humanity to be tracked, and the most pressing issues to be identified. The early identification of such issues is relevant for researchers, policy-makers, and the general public, providing an opportunity to consider what anticipatory or future action might need to be taken.

Environmental hazards associated with the global food system threaten societal integrity. Yet, there is a major data gap in the global understanding of how the prevalence of hazards is changing over time, how different classes of hazard are distributed, and whether the combined literature represents hazard prevalence equitably across research, policy and legislation, and news. Here, we explore this data gap, leveraging global research, policy, and news databases. We reveal increasing attention on food system hazards over time, in line with major geopolitical events. Coverage on environmental hazards is not distributed equally geographically, and media attention does not match research and policy evidence focus. Climate change and water scarcity in particular receive substantial attention across all source types, whilst, for example biodiversity loss, genetic erosion, or harmful algal blooms receive much less. Environmental, financial and food systems sustainability damage due to hazard neglect should be avoided and a first step is to understand, map, and quantify biases in focus.