Dr Jacob Webber, a researcher in the School of Informatics at the University of Edinburgh, has been awarded a UK Intelligence Community Postdoctoral Research Fellowship from the Royal Academy of Engineering. The fellowship will support a two-year research programme exploring new ways to identify AI-generated speech, commonly known as audio deepfakes.Audio deepfakes can be used to impersonate individuals, spread misinformation, commit fraud, or undermine trust in digital communications. As the technology behind synthetic speech rapidly improves, detecting convincing deepfake audio has become an increasingly important challenge. Detecting deepfakes by understanding real speech Dr Webber's fellowship project, Explainable acoustic profiles of speech deepfakes, takes a different approach to many existing deepfake detection systems.Rather than training systems to recognise examples of known deepfakes, the research will focus on modelling the characteristics of authentic human speech itself.The project will draw on the physical and linguistic properties of natural speech, including features such as accent consistency and vocal tract behaviour, to develop AI systems capable of distinguishing genuine speech from synthetic imitations.By learning the characteristics of real human speech, the approach aims to identify previously unseen forms of deepfake audio while also providing explanations for why a recording has been flagged as suspicious. Growing challenge of synthetic media The rapid development of generative AI tools has made synthetic audio significantly easier to produce. While these technologies have many legitimate applications, they also present challenges for security, misinformation and digital trust.According to the Royal Academy of Engineering, existing deepfake detection techniques often rely on known examples of manipulated audio, limiting their ability to identify new and emerging forms of synthetic speech.Dr Webber's research aims to address that challenge by developing methods that do not depend on previously encountered deepfakes, making them more resilient as the technology evolves. Supporting the next generation of research leaders The UK Intelligence Community Postdoctoral Research Fellowships support early-career researchers undertaking fundamental research in areas of interest to the UK's intelligence, security and defence communities.The fellowships are jointly supported by the Royal Academy of Engineering, UK Research and Innovation (UKRI) and the Government Office for Science.Announcing this year's awards, the Royal Academy of Engineering highlighted the importance of supporting outstanding researchers working on challenges ranging from clean energy and healthcare to artificial intelligence and national security. Further reading Royal Academy of Engineering fellowship announcement Dr Jacob Webber fellowship profile This article was published on Wednesday 21 October 2026
