The Alan Turing Institute is the United Kingdom's national institute for data science and artificial intelligence, founded in 2015. It is named after Alan Turing, the British mathematician and computing pioneer.
Recent advances in machine learning and AI, including Generative AI and LLMs, are disrupting technological innovation, product development, and society as a whole. AI's contribution to technology can come from multiple approaches that require access to large training data sets and clear performance evaluation criteria, ranging from pattern recognition and classification to generative models. Yet, AI has contributed less to fundamental science in part because large data sets of high-quality data for scientific practice and model discovery are more difficult to access. Generative AI, in general, and Large Language Models in particular, may represent an opportunity to augment and accelerate the scientific discovery of fundamental deep science with quantitative models. Here we explore and investigate aspects of an AI-driven, automated, closed-loop approach to scientific discovery, including self-driven hypothesis generation and open-ended autonomous exploration of the hypothesis space. Integrating AI-driven automation into the practice of science would mitigate current problems, including the replication of findings, systematic production of data, and ultimately democratisation of the scientific process. Realising these possibilities requires a vision for augmented AI coupled with a diversity of AI approaches able to deal with fundamental aspects of causality analysis and model discovery while enabling unbiased search across the space of putative explanations. These advances hold the promise to unleash AI's potential for searching and discovering the fundamental structure of our world beyond what human scientists have been able to achieve. Such a vision would push the boundaries of new fundamental science rather than automatize current workflows and instead open doors for technological innovation to tackle some of the greatest challenges facing humanity today.
User-facing ‘platform safety technology’ encompasses an array of tools offered by social media platforms to help people protect themselves from harm, for example allowing people to report content or block other users. These tools are an increasingly important part of online safety; however, little is known about how users engage with them. We present findings from a nationally representative survey of UK adults examining their experiences with online harms and safety technologies. The results show that online harm is widespread: 67% of respondents report having encountered harmful content online. Among those who are aware of safety tools, over 80% have used at least one, indicating high uptake when knowledge of the tools is present. Awareness of specific tools is varied, with people more aware of ‘post hoc’ safety tools, taken in response to harm exposure (such as reporting or blocking), than preventive measures (such as altering feed algorithms). However, satisfaction with safety technologies is generally low. People who have previously seen online harms are more likely to use safety tools, implying a ‘learning the hard way’ route to engagement. Those higher in digital literacy are also more likely to use some of these tools, raising concerns about the accessibility of these technologies. In addition, women are more likely to engage in particular types of online ‘safety work’. These findings have significant implications for platform designers, regulators, researchers and policymakers seeking to create a safer and more equitable online environment.
Cooperative autonomous robotic systems have significant potential for executing complex multi-task missions across space, air, ground, and maritime domains. But they commonly operate in remote, dynamic and hazardous environments, requiring rapid in-mission adaptation without relying on fragile or slow communication links to centralized compute. Fast, on-board replanning algorithms are therefore essential to enhance resilience for these systems, but do not yet exist. Reinforcement Learning (RL) shows strong promise for efficiently solving mission planning tasks formulated as Travelling Salesperson Problems (TSPs), but existing methods: 1) are unsuitable for replanning, where agents do not start at a single location; 2) do not allow cooperation between agents; 3) are unable to model tasks with variable durations; or 4) lack practical considerations for on-board deployment. Here we address this gap by defining the Cooperative Mission Replanning Problem as a novel adaptation of multiple TSP, and develop a new encoder/decoder-based RL model to solve it effectively and efficiently. Using a simple example of cooperative drones, we show our replanner consistently (90% of the time) maintains performance within 10% of the state-of-the-art LKH3 heuristic solver, whilst running 85-370 times faster on a Raspberry Pi. This work paves the way for increased resilience in autonomous multi-agent systems.
Emergent communication enables agents to develop bespoke languages that improve communication efficiency. Despite the known importance of temporal structure in natural language, there is no existing evidence of temporal references in emergent communication. This paper addresses this gap, by exploring how agents communicate about temporal relationships. We analyse three potential factors for the emergence of temporal references: environmental, external, and architectural. Our experiments demonstrate that altering the loss function is insufficient for temporal references to emerge; rather, architectural changes are necessary. A minimal change in agent architecture, using a different batching method, allows the emergence of temporal references. This modified design is compared with the standard architecture in a temporal referential games environment, which emphasises temporal relationships. The analysis shows that over 95% of the agents with the modified batching method develop temporal references, without changes to their loss function. We consider temporal referencing necessary for future improvements to the agents’ communication efficiency, enabling future agents to use a closer to optimal coding as compared to purely compositional languages. These insights provide the basis for incorporation of temporal references into other emergent communication settings, and investigation of other aspects of language.
This study presents a framework for automating the generation and validation of machine-readable vulnerability statements, known as Vulnerability-Exploitability Exchange (VEX) artefacts, within secure research environments. The work addresses a critical limitation in existing vulnerability reporting, where static scoring systems often fail to capture whether a flaw is truly exploitable in a specific analytic context. By integrating structured metadata capture, runtime instrumentation, and cryptographically verifiable provenance, the framework classifies vulnerabilities as affected, fixed, or not relevant, supported by machine-readable evidence bundles. The methodology was evaluated using containerised applications seeded with deliberately vulnerable components. Software bill of materials and vulnerability scanners were applied to generate baseline inventories, while reproducibility frameworks validated that results could be independently replicated. Findings demonstrate that automated VEX generation can reduce false positives by distinguishing theoretical from actionable risks, thereby improving security assurance and reproducibility in federated infrastructures. At the same time, the research acknowledges significant challenges. Computational overhead from multi-layered monitoring, dependence on external tools, and the risk of false negatives introduce barriers to adoption. Broader pilot studies across heterogeneous domains and benchmarking on standardised testbeds are required to enhance generalisability. Privacy concerns from extensive runtime monitoring and the need for sustainable maintenance models also demand attention. By combining automation with human oversight and aligning with emerging standards, the study contributes a reproducible, auditable, and context-sensitive approach to vulnerability management. The work provides both a proof-of-concept and a roadmap for refining security practices in sensitive computational environments.
