34028 research outputs found
Sort by
Food Standards Agency Science Council Report of Project ‘Artificial Intelligence Applications in Food Safety and Authenticity'
No full textArtificial Intelligence (AI) technologies are advancing rapidly with significant potential to transform how food is produced, managed and regulated. In the food system, AI offers the prospect of more efficient, predictive and responsive assurance processes, ranging from real-time detection of hazards to automated documentation checks and predictive modelling of risks. At the same time, the adoption of AI raises important questions about accountability, transparency and trust, particularly where automated systems interact with actions to comply with the Food Safety Act (1990) and the due diligence defence relied upon by Food Business Operators (FBOs).The Food Standards Agency (FSA) has a responsibility to ensure that innovation does not undermine consumer protection, regulatory oversight, or public confidence. This Science Council project was established to explore the likely applications of AI in food safety and assurance; identify the benefits and risks and consider implications for the FSA’s role as a regulator. The study drew on academic and policy evidence with key insights from a June 2025 workshop attended by food businesses, regulators, assurance providers, academics and technology developers. Four case studies were used to anchor discussions in realistic scenarios where AI (all forms from machine learning applications to generative and emerging AI systems) is likely to be applied now and in the near future:AI-driven safety and regulatory compliance evaluation for manufactured foodsAI-supported data pack generation for third-party certification and assuranceAI-assisted detection of infections and other pre/post-mortem pathologies in UK abattoirsAI-powered document inspection at UK ports of entryThese case studies, and post-workshop discussions amongst the project team, wider science council and FSA staff, highlighted both opportunities and challenges. AI could enable faster detection of hazards, more consistent and scaled inspections, wider surveillance across supply chains and real-time data analytics to help target interventions. It could reduce reliance on sampling or retrospective checks and free inspectors or auditors to focus on higher-value tasks. However, workshop participants also identified risks: AI systems may embed bias or drift if not carefully validated; they can generate outputs that are difficult to explain or reproduce; and in poorly managed businesses, they could conceal weaknesses behind apparently robust documentation. Across all scenarios, the need for human oversight, clear accountability, explainability of decisions and robust validation of AI tools was consistently emphasised. Especially in these early stages of AI deployment, when both industry and regulators are still learning how AI systems can be safely applied, vigilance is essential; over time, experience will help clarify the contexts in which AI delivers most value and the necessary safeguards.For the FSA, the implications are clear. AI has the potential to strengthen assurance processes, but only if deployed within strong governance frameworks and supported by clear guidance. We recommend the Agency should clarify best practice for the integration of AI within FBO accountability; continue its promotion of data standards and sharing; provide guidance for FBOs on responsible use; and work with industry, standards bodies, and other regulators to support codes of practice and validation mechanisms. At the same time, it must remain alert to the risks of hype and over-reliance, ensuring that AI enhances, rather than displaces, the human accountability that underpins food law.Due to the rapid emergence of AI technologies, and notwithstanding the present limited case study evidence from scaled industrial use, there was broad consensus that existing UK food safety regulations are sufficiently robust to encompass the use of currently known AI systems in the food system. This study does not therefore call for immediate changes to legislation. However, the FSA will need to continually monitor developments in AI, assess their impacts on assurance processes and remain prepared to act if gaps emerge. Future regulatory attention may be required in areas such as validation standards, data governance, or liability frameworks should AI adoption accelerate, or if new classes of tools present novel risks.The following recommendations aim to provide the FSA with practical steps to support safe and responsible AI adoption, ensuring that innovation contributes to a more predictive, preventative, and trusted food safety system.</p
Passive outperforms active whisking in a deep learning shape classification task
No full textTactile sensing with robotic whiskers has been shown to enable shape and texture recognition. Biological studies and robot practicalities -- avoiding breaking whiskers -- strongly suggest active whisking (i.e. retracting the whisker shortly after contact) to be preferable to passive (i.e. open-loop whisking patterns during contact). While active whisking may protect whisker sensors, the effect on accuracy has not been studied in detail, though there are some early indications using suboptimal, pre-deep learning classifiers that it may degrade accuracy. We perform a controlled study of active vs passive whisker sensing for shape recognition, using multiple current deep learning models. Experimental results show that passive whisking outperforms active classification performance by 10\%, with an F1-score of 0.85 on the passive whisking dataset across various shape categories, compared to 0.74 for active whisking. This raises questions about how to trade off accuracy for physical whisker protection. Our experiments are the first to be performed using a completely open source hardware and software stack, so they can be used as a replicable baseline to enable future metric-driven incremental improvement whisker sensing research, as in mature fields such as machine vision.</p
The influence of stigma and resilience on attitudes toward mental health help-seeking in athletes from the United Kingdom
No full textAthletes have been identified as a reluctant group to seek mental health support and often report stigma and negative attitudes toward help-seeking as barriers. Resilience is seen as a desired psychological characteristic by those involved in sport, shown to guard against stressors, and related with enhanced psychological wellbeing. Researchers, however, have yet to investigate the relationship between stigma and resilience, and how each may be associated with attitudes toward mental health help-seeking in athletes. A total of 297 athletes from the United Kingdom completed an online survey containing measures of personal stigma, self-stigma, resilience, and positive attitudes toward mental health help-seeking. Correlations showed both stigma types were positively related to one another but had no association with resilience. A multiple linear moderation model further revealed both stigmas were negatively related to positive help-seeking attitudes, whereas resilience had no association. No moderation effects were found between any pair of independent variables.</p
Training human super-recognizers’ detection and discrimination of AI-generated faces
No full textGenerative adversarial networks (GANs) can create realistic synthetic faces, which have the potential to be used for nefarious purposes. The synthetic faces produced by GANs are difficult to detect and are often judged to be more realistic than real faces. Training programmes have been developed to improve human synthetic face detection accuracy, with mixed results. Here, we investigate synthetic face detection and discrimination in super-recognizers (SRs; who have exceptional face recognition skills), and typical-ability control participants. We also devised a training procedure which sought to highlight rendering artefacts. In two different experimental designs, we found that SRs (total N = 283) were better at detecting and discriminating synthetic faces than controls (total N = 381), where control participants were below chance without training. Trained SRs and controls had significantly better performance than those without training, and the magnitude of the training effect was similar in both groups. Our results suggest that SRs are using cues unrelated to rendering artefacts to detect and discriminate synthetic faces, and that an easily implementable training procedure increases their performance to above chance levels. These results have implications for real-world scenarios, where trained SRs' performance could be harnessed for synthetic face detection.</p
Assessing barriers to blockchain technology adoption in food supply chains: a rough DEMATEL analysis
No full textThis study investigates the key obstacles to the adoption of blockchain technology in the food sector. It identifies technological, organisational and regulatory barriers that hinder implementation, and explores how firms in food supply chains perceive and react to these challenges. The findings provide a framework for managers and policymakers to address adoption inhibitors and steer blockchain-enabled innovation in food systems.Purpose: This paper identifies and prioritises the prominent barriers to blockchain technology (BCT) adoption in food supply chains (FSCs) and the interrelationships among these barriers and demonstrates how integrating BCT can enhance sustainability, traceability, transparency and reliability.Design/methodology/approach: A pairwise comparison survey of academics and practitioners with expertise in FSCs/BCT was conducted, then application of the Decision-Making Trial and Evaluation Laboratory (DEMATEL) technique modelled and quantified the causal links between the barriers. DEMATEL's ability to identify interdependencies and feedback loops provided a detailed understanding of how variables interact in a cause-and-effect network.Findings: Key barriers to the adoption of BCT in FSCs included collaboration, coordination and communication in FSCs, the cost of BCT implementation and BCT customisation costs, all of which were interrelated and affected the availability of tools and standards. Overall, the cost of bespoke design and customisation of BCT systems was perceived as the most influential barrier, while lack of access to legal permissions for cryptocurrency use was considered to have the least impact.Practical implications: Successful BCT adoption depends on addressing implementation costs, enhancing collaboration, coordination and communication in FSCs and developing robust tools and standards that are accepted by all stakeholders.Originality/value: This study explores the adoption of BCT in FSCs using data from two key stakeholder groups: academics and practitioners, demonstrating how coupling rough set theory with DEMATEL reduces subjectivity and creating a robust framework for mapping causal relationships among influencing barriers. Moreover, the work bridges to future research by identifying further areas for research.</p
The Importance of Feedstock and Process Control on the Composition of Recovered Carbon Black
No full textPyrolysis has emerged as a commercially viable material recovery process capable of supporting circularity in the tyre industry. Here it is demonstrated that a high degree of control can be imparted over the UK tyre waste stream and that statistically different feedstocks can be used to produce different grades of rCB based on their ash contents. The lower ash content rCB produced from truck tyres had superior in-rubber properties, closely matching those of the N550 reference. Silica, absent a coupling agent, is known to be less reinforcing than CB, lowering the reinforcing behaviour of the high ash content rCB variant produced from car tyres. This justifiably places ash content within the classification and specification development discussion. However, proximate analysis of UK waste tyres suggests that typical rCB ash specifications of <20 wt% are unrealistic. Such limits would force producers to consider modifying process conditions to allow the deposition of carbonaceous residues to artificially dilute the ash content. This study investigated this process philosophy but conclusively demonstrated that carbonaceous residue is more detrimental to rCB performance than ash content. As such, carbonaceous residue content demands far more attention from the industry than it is currently afforded.</p
Investigating Federated Learning Approaches to Intrusion Detection Systems on IoT Devices
No full textThe rapid growth in Internet of Things (IoT) devices has introduced significant se?curity challenges as these systems present an increasingly multifaceted attack surface for threat actors. Existing intrusion detection systems (IDS) for IoT devices focus almost exclusively on the machine learning performance of these devices while over?looking the practical challenges unique to an IoT deployment. This thesis proposes a novel approach to federated transfer learning and federated isolation forest, aiming to make lightweight models which retain all of the performance of standard feder?ated anomaly detection algorithms. These models have been trained across various network sizes with a partitioned global dataset. Experiments on the CIC-IoT2023 dataset demonstrate that certain federated networks with the partitioned global set approach can remain performant up to 1024 nodes, with room for future work and optimisations. Federated isolation forest in particular shows strong potential for future research. This work highlights the viability of federated learning in securing the next generation of IoT devices against IoT-specific cyber attacks, and explores ap?proaches to federated learning that can be applied to other domains within computer science.</p
'I felt as though I'd been in Jail': Women's experiences of maternity care during labour, delivery and the immediate postpartum
No full textIt has been widely recognized, both in the UK and internationally, that there is a need for a multidimensional or holistic approach to maternity care, which incorporates psychological as well as physical aspects, in order to optimize women's experiences both in the intra- and postpartum period. Central to such an approach is the relationship between women and maternity care staff. The aim of this study was to explore the impact of maternity care staff on women's experiences, and feelings associated with the childbirth process. Semi-structured interviews were conducted with 24 primiparous and multiparous women, and transcripts analysed using open and axial coding with triangulation. Three main themes emerged from women's accounts: perceptions of control, staff attitudes and behaviours, and resource issues. Each of these themes was evident throughout the various stages of the childbirth process, in the delivery suite, on the maternity ward, and specifically in relation to breastfeeding. In the women's accounts, feelings of little control were related to inadequate information provision, poor communication, and no opportunity to influence decision making. These, together with the negative attitudes and behaviours of maternity staff, and issues of under-resourcing, were often linked to negative feelings such as fear, anger, disappointment, distress, guilt, and inadequacy. These findings illustrate the importance of maternity care staff recognizing women's psychological and emotional needs during the childbirth process, and the impact that they themselves may have on women's experiences. These issues are discussed with reference to the wider debate on authority and power within the medical relationship, from a feminist viewpoint. © 2005 SAGE.</p
Using Role Substitution to Improve Oral Health in Care Homes: A Process Evaluation
No full textBackground: The oral health of many older adults residing in care homes is poor and service provision is limited. Role substitution has been suggested as a potential model to improve service provision in this context and describes the reallocation of tasks from a dentist to other members of the dental team. Objectives: To undertake a theoretically informed process evaluation alongside a pragmatic cluster-randomised controlled trial to determine whether the use of Dental Therapists and Dental Nurses could improve the oral health of dependent older adults in care homes in the UK. Materials and Methods: Semistructured interviews were held with 17 key stakeholders responsible for intervention delivery. Parallel observations were utilised during the intervention delivery phase in 22 homes. Both were conducted inductively using the main themes from the Promoting Action on Research Implementation in Health Services (PARIHS) framework to focus on intervention delivery and implementation. Results: Stakeholders were receptive to the potential of using role substitution in this setting and saw this as a viable alternative to current practice. Partnership working was considered key, but was not always observed, and some care staff did not see oral health as their responsibility. The physical environment of the care home setting created a number of challenges, and sugary food and drinks were ubiquitous and formed an important part of the day-to-day structure within the home. Conclusion: Although role substitution has the potential to meet the needs of dependent older people, a number of challenges exist in promoting oral health and delivering service provision. © 2025 The Author(s). Gerodontology published by Gerodontology Association and John Wiley & Sons Ltd.</p
Your Teeth, You Are in Control: A Process Evaluation of the Implementation of a Cognitive Behavioural Therapy Intervention for Reducing Child Dental Anxiety
No full textAim: To explore the views of patients, caregivers, and dental professionals on the factors that influence implementation, processes, and effectiveness of a guided self-help cognitive behavioural therapy (CBT) intervention, ‘Your teeth, you are in control’ (YTYAIC), in the CALM trial. Methods: Semi-structured interviews were conducted as part of this qualitative component of the process evaluation, and data were analysed using a framework approach based on the Consolidated Framework for Implementation Research (CFIR) and the Five Areas Model of CBT. Results: Thirty-seven participants were recruited. Potential mechanisms of action were identified using the Five Areas Model of CBT. Participants felt the intervention may exert change through targeting unhelpful thoughts and feelings (e.g., building trust and perceptions of control) and behaviours (e.g., encouraging effective communication and coping strategies) and facilitating a more positive situational context (e.g., developing more supportive relationships). Enablers (e.g., adaptability, design and delivery) and barriers (e.g., time/resource constraints, cost) to implementation were identified using the CFIR. Conclusions: This study revealed multiple potential mechanisms of action which could reduce dental anxiety and examined how implementation and contextual factors may influence this change process. The results of the research revealed that the intervention could be implemented in primary dental care and identified the potential barriers which should be addressed to aid successful implementation of the intervention in real world contexts. Trial Registration: This clinical trial has been registered with an international registry and has been allocated an International Standard Randomised Controlled Trial Number (ISRCTN27579420). © 2025 The Author(s). Community Dentistry and Oral Epidemiology published by John Wiley & Sons Ltd.</p