Project Abstract

With artificial intelligence (AI) now deeply embedded across various sectors, ensuring the security of AI systems from potential threats is paramount. This project proposes an investigation into methodologies for safeguarding AI, focusing primarily on identifying and mitigating security threats that could compromise AI systems. The growing dependency on AI for critical operations across industries highlights the urgency for a robust security framework that not only prevents data breaches but also shields the AI from manipulative disruptions. The study will systematically review the existing literature to outline the current state of AI vulnerabilities and the typical avenues through which these systems are exploited. The project will analyse insights from diverse academic sources and security reports to develop a tailored framework of best practices and preventive measures, strengthening AI’s defences against known and emerging threats. This will include recommendations for continuous security assessments and the implementation of advanced cryptographic measures, which assumes that breaches are not only possible but likely, thereby necessitating constant vigilance and verification within the AI operational environment. The contribution of this research will focus on developing a comprehensive model and identifying appropriate tools for the security assurance of AI systems. The project aims to explore and evaluate various security tools and techniques, determining their effectiveness in real-world scenarios. This effort will provide a toolkit that can be utilized by organizations to assess, enhance, and maintain the security of their AI systems effectively. The goal of this effort is to critically analyse both the advantages of AI in automating and optimizing operations, and the cons, notably the security gaps that leave AI systems vulnerable to attacks. By addressing these vulnerabilities comprehensively, the project will offer a vision for a more secure AI landscape, underpinned by rigorous research and tailored security strategies.

Keywords: Cyber Security, Artificial Intelligence (AI), Security of AI

 Conference Details

Session: Presentation Stream 26 at Presentation Slot 1

Location: GH043 at Wednesday 8th 13:30 – 17:00

Markers: George Brooks (GTA), Markus Roggenbach

Course: MSc Cyber Security, Masters PG

Future Plans: I’m looking for an industry placement