MSc Computer Science - PROM05: My Research

Field of Study and Project Overview

The research project for my dissertation is based in the fields of AI and cybersecurity, with a specific focus on the detection of deepfake media on social networking platforms. Advances in generative AI have made it increasingly easy to create realistic fake images, videos, and audio, raising serious concerns around misinformation, privacy, and public trust. As a result of this, developing reliable and accessible detection tools has become an important area of study.

For my project proposal, I focused on the concept of a prototype deepfake detection tool for use within a social media environment. Rather than relying solely on cloud-based systems, the proposal explored the use of a lightweight, client-side solution that could analyse media directly on user devices. This approach aims to reduce latency, improve user privacy, and give individuals greater control over how their data is processed.

The proposed solution is intended to monitor visual content, including images and short-form video, as these are commonly used in online misinformation. The aim is that it will apply machine learning techniques to identify signs of manipulation and present the results through clear visual indicators and context-based warnings. Instead of automatically removing content, the tool will be designed to support informed user decision-making.

My proposal also discussed several practical constraints, including limited processing power on client devices and the difficulty of detecting low-quality or intentionally modified deepfakes. These challenges highlight the importance of balancing technical accuracy with usability and performance.

To address these issues, the research design combines technical testing with user-focused evaluation. This allows the project to assess not only how accurately deepfakes can be detected, but also how effective and trustworthy the system feels in real-world use.

There will be many blog entries like this, and I will build upon the original proposal by documenting the development of the prototype, reflecting on key challenges, and exploring how current research continues to influence deepfake detection technologies. Together, these entries will demonstrate how the project evolves from initial design to practical implementation.

References

The references below represent the foundational academic and industry research that has guided the creation of my research questions, methodology, and project deliverables.

Artificial Intelligence Review (2024). Deepfake video detection: Challenges and opportunities. Artificial Intelligence Review. Available at: https://link.springer.com/article/10.1007/s10462-024-10810-6 (Accessed 4 Jan. 25).

Babaei, R., Cheng, S., Duan, R. and Zhao, S. (2025). Generative artificial intelligence and the evolving challenge of deepfake detection: A systematic analysis. Journal of Sensor and Actuator Networks, 14(1), p. 17. Available at: https://www.mdpi.com/2224-2708/14/1/17 (Accessed 4 Jan. 25).

Nguyen, T.T., Nguyen, C.M., Nguyen, D.T. et al. (2023). Deep learning for deepfakes creation and detection: A survey. IEEE Transactions on Neural Networks and Learning Systems, 34(9), pp. 4553–4571. Available at: https://arxiv.org/abs/1909.11573 (Accessed 4 Jan. 25).

Qureshi, S.M., Saeed, A., Almotiri, S.H., Ahmad, F. and Ghamdi, M.A.A. (2024). Deepfake forensics: A survey of digital forensic methods for multimodal deepfake identification on social media. PeerJ Computer Science. Available at: https://pubmed.ncbi.nlm.nih.gov/38855214/ (Accessed 4 Jan. 25).