University of Greenwich

Phishing attacks continue to grow in scale and sophistication, causing substantial financial losses and privacy breaches worldwide. Recent advances in large language models (LLMs) have brought significant changes to the generation and detection of phishing content. This study systematically investigates the dual role of LLMs in facilitating phishing attacks and strengthening countermeasures. Using the PRISMA methodology, authors screened 142 records published between January 2023 and April 2025 and identified 36 eligible studies from major academic databases, including IEEE Xplore, ScienceDirect, ACM Digital Library, Web of Science, and Scopus. A comprehensive and rigorous analysis was conducted of research trends/themes over time, dataset characteristics, and the LLM architectures/models employed. The findings reveal that most studies relied on manually generated datasets rather than publicly available benchmark datasets, and that GPT-based models received considerably more attention than other LLM architectures. The review demonstrates that LLMs substantially enhance the generation of phishing content by producing coherent, contextually relevant, and persuasive email and website content. This capability lowers the technical barrier for attackers and potentially increases attack effectiveness. Conversely, LLMs also strengthen defensive strategies by enabling more effective analysis of textual and visual content for phishing detection. In many cases, LLM-based approaches outperform traditional machine learning and deep learning methods and, in certain contexts, approach or match human-level performance. Overall, the findings suggest that LLMs have accelerated and automated phishing-related processes, simultaneously intensifying the threat landscape and advancing defensive capabilities. • The first in-depth study to review LLMs usage in Phishing attack generation and detection. • The study reveals LLMs have accelerated and automated phishing-related processes, elevating both threats and defence mechanisms. • GenAI-based multimodal phishing attacks are on the rise due to the wider adoption of GenAI tools in general.

Underutilized legumes such as Bambara groundnut, Lima bean, African locust bean, and Black gram offer affordable sustainable protein sources, yet remain insufficiently characterized for ingredient development. Nixtamalization is established in cereals for enhancing mineral bioavailability, but its effects on legumes are poorly understood. Saccharification, intended to hydrolyze carbohydrates and yield protein-rich powders, lacks systematic evaluation in many legumes. This study applied nixtamalization followed by saccharification to legume flours and conducted comprehensive characterization using the Dumas method for protein quantification and SEM-EDX, XRD, and FT-IR for microstructural and compositional analysis. Observations of intact starch granules embedded within protein bodies indicated the absence of gelatinization; XRD confirmed amorphous powder profiles; FT-IR identified functional groups consistent with carbohydrates, proteins, and lipids; and SEM-EDX detected multiple elements (Ca, K, Mg, Cu). Nixtamalization consistently increased calcium content, demonstrating its utility for mineral enhancement in legume matrices. In contrast, saccharification did not uniformly improve protein concentration: significant decreases in protein were observed in Bambara groundnut and Lima bean, while African locust bean and Black gram remained stable. These findings delineate processing-composition interactions that are legume-specific and advance the valorization of underutilized legumes. In future optimizing of process parameters to develop legume-derived protein concentrates and functional ingredients aligned with sustainable nutrition and diversified protein supply will be essential . • Nixtamalization was applied to underutilized legumes to enhance protein usability. • Nixtamalization enhanced Ca content in underutilized legume powders. • α-Amylase saccharification modified legume microstructure and protein matrix. • XRD showed amorphous protein-starch structures in treated legume powders. • Single-enzyme saccharification reduced, rather than concentrated, protein.

Cassava brown streak disease (CBSD), caused by cassava brown streak virus (CBSV; Ipomovirus brunusmanihotis) and Ugandan cassava brown streak virus (UCBSV; Ipomovirus manihotis) (family Potyviridae, genus Ipomovirus), is increasingly becoming a threat to cassava production in several parts of Africa, especially in Eastern, Central and Southern Africa. In Kenya, the disease continues to wreak havoc on cassava production leading to a significant reduction in crop yields and economic losses of up to USD 1 billion. Variation in virus populations make the control of CBSD challenging as virus genomic variation can affect the accuracy of diagnostic tests, lead to resistance breaking isolates and jeopardize strategies of breeding for resistance. CBSV and UCBSV populations obtained from cassava fields in Kenya were characterized. In total, 44 new complete sequences of CBSV and UCBSV were assembled and 40 sequences successfully submitted to GenBank. Single Nucleotide Polymorphism (SNP) analysis revealed that the cylindrical inclusion protein (CI) is the most stable region across the genome of CBSV and UCBSV. In contrast, protein 1 (PI) and the coat protein (CP) were the most hypervariable regions. Phylogenetic analysis showed three major geographical groupings for both UCBSV and CBSV isolates, suggesting a continued spread of the viruses through human-mediated movement of infected planting materials. The data obtained in this study can support the development of disease management strategies through improved molecular diagnostic tests and targets for breeding for resistance against CBSD.

Engaging citizens in biodiversity and nature-focused decision-making processes has been limited in scope and design to date. Increasing polarisation is occurring, especially between resource users and environmental activists on questions of biodiversity protection and restoration, while latent support for biodiversity is not well quantified, but appears much higher than publics are aware of. There is extensive experience developing of facilitating living lab approaches to innovation and problem solving, and other dialogic approaches to multi-actor learning are emerging that have a focus on place-based restoration and protection of nature and can inform spatial planning. Such transdisciplinary approaches are complex in practice, and struggle to resolve or overcome power inequalities. We will reflect upon methodological innovations and author's fieldwork experiences in facilitating living labs and social learning processes in contexts of complex power dynamics. Deliberative approaches have been applied more widely to other topics, such as climate change, rather than on biodiversity, nature and societal futures in specific places. A notable exception being the Irish Biodiversity Citizen Assembly. More experimentation is required in deliberation to give space to plural ontologies, epistemologies and practices, as well as increased investment in the evidencing outcomes. Finally, there are newly emerging creative approaches that engage the arts to provoke participants in terms of challenging dominant hegemonic assumptions and thereby to generate rich new future possibilities. Across all these spaces of citizen engagement in decision-making on biodiversity, nature and societal futures in contexts of climate change, the rise of AI creates new possibilities and raises new ethical concerns. This paper offers a typology of citizen engagement approaches, explores a range of cases cutting across these approaches to distil key insights and lessons on justice, power and plurality, and the potential of these deliberative, dialogic and creative approaches to citizen engagement for sustainable nature and people futures. We distil key insights from emerging evidence and outline new directions and challenges including with respect to other-than-human as well as human agency and values.

This paper explores the integration of international perspectives into forensic science education, emphasising the pedagogical value of global engagement and authentic learning experiences. Drawing on constructivist theories and first-hand experiences in countries such as Nepal, I argue that exposure to diverse forensic practices enhances student understanding of legal frameworks, cultural contexts and scientific adaptability. I advocate a balanced, evidence-based approach that fosters critical thinking, inclusivity and global awareness in forensic science pedagogy.

Globally in 2021, about 1 in 4 children and young people (CYP) below age 25 lived with poor mental health. Various forms of digital health interventions (DHIs) have been introduced in attempts to address unmet mental healthcare needs among young people. DHIs use websites or applications in computers, tablets, or smartphones to deliver mental health interventions, which are either self-directed, or therapist guided. This rapid review synthesises the evidence on the cost-effectiveness of DHIs for supporting mental health in CYP. A literature search was carried out in MEDLINE and PsycINFO using Ovid platform. The search was restricted to peer-reviewed studies published in English between Jan 2018 to May 2025. Eligible studies were identified using the population, intervention, comparator, outcome, and study design (PICOS) framework. Studies were eligible for inclusion if they provided a trial- or model-based full economic evaluation or return on investment of a DHI targeting mental health in CYP up to 25 years of age. Titles and abstracts of 1,265 records were screened. 15 studies met the criteria for inclusion following full-text screening of 28 studies. The focal outcomes in the studies included reduction in symptoms linked to depression, anxiety, and alcohol use. Most studies evaluated computerised cognitive behaviour therapy (CBT) or interventions that included aspects of CBT as the primary intervention. All studies except two reported the DHIs to be cost-effective, at least under certain conditions. Two model-based and three trial-based studies reported that the interventions were dominant (cost saving and more effective) with respect to at least one outcome measure. There is some evidence to suggest DHIs hold the potential to expand and extend mental health support for CYP with minimal therapist involvement. Yet, the evidence is not conclusive due to short follow-up periods, variability in the methodological approaches and reporting of results. High quality evidence on cost-effectiveness of DHIs with comparable methodological approaches is needed to inform implementation decisions.

The intramembrane 'rhomboid' protease family is almost ubiquitous across evolution, with its well-conserved transmembrane domains typified in crystal structures of bacterial representatives, such as the Escherichia coli GlpG. In contrast with accumulating data on rhomboid function in higher organisms, where roles in human disease are an incentive for study, findings remain sparse about the functions and substrates of the prokaryotic enzymes, even though these provided the earliest protein structures. In particular, nothing at all is known about the rhomboid proteases of photosynthetic prokaryotes despite the importance of cyanobacteria as relatives of the progenitor of chloroplasts. Findings relating to the cyanobacterial enzymes would complement data on plant plastid rhomboids from work in Arabidopsis thaliana. Synechocystis sp. PCC 6803 was used, therefore, to investigate conserved photosynthetic functions across evolution for this protein family. Reverse-genetics studies using Slr1461, the single rhomboid protease of Synechocystis 6803, did not reveal a non-photochemical quenching phenotype as observed for the Arabidopsis RBL10 null mutant, which lacked a chloroplast outer membrane rhomboid. The Slr1461 mutant exhibited a marginal change in pigment composition, and its growth rate was only slightly different from that of WT under optimal light intensity. The most dramatic effect of the inactivation of Slr1461 was the mutant's distinct inability to reduce photosynthetic activity under mixotrophic conditions. Concurrent phototrophy and heterotrophy in mixotrophic growth aids survival and competitiveness in phytoplankton, allowing conservation of energy by reducing the need for uptake and fixing of CO2 when an organic carbon source is available. It was notable, therefore, that, in the absence of the Slr1461 rhomboid, the steady-state mRNA levels were reduced for a subset of genes encoding facilitators of high-affinity CO2 import and of transcriptional regulators of the carbon-concentrating mechanism (CCM). Slr1461 activity was also linked with that of another membrane protease, the AAA protease FtsH2, which was likewise observed to act within regulatory networks for the cyanobacterial carbon uptake mechanism. Aberrant transcript levels were most evident specifically under high CO2 conditions, when the impact of Slr1461 enzymatic activity appeared to be upstream of NdhR, a central, controlling transcription factor of the CCM.

Abstract In response to the pressing challenge of global climate change, advancing a low-carbon energy transition has emerged as a key international priority. As an integral policy instrument to guide this transition, carbon pricing is increasingly adopted by countries and regions worldwide. Drawing on a spatial panel model and covering 115 countries, this study investigates the effects of carbon pricing on carbon emission reduction and compares the outcomes between single and composite carbon pricing instruments. The spatial spillover effects of carbon pricing policies exhibit multidimensional heterogeneity. Hybrid carbon pricing policies form a cross-regional emission reduction network through regional synergistic governance mechanisms. In contrast, carbon tax and emissions trading systems (ETS) are associated with the ‘pollution paradise hypothesis’ and the ‘race to the bottom effect’, respectively. Further, it elucidates how different carbon pricing policies leverage unique economic and energy-related mechanisms to facilitate emission abatement. The findings offer important insights for policymakers aiming to optimize carbon pricing schemes that effectively support the global energy transition.