Artificial intelligence is revolutionising cyberspace on a daily basis, providing new opportunities for both cyberattacks and cyber defence. It’s easier than ever for digital criminals to access innovative technologies and use them to breach once-robust security systems, often just by automating their attempts. Public data, vulnerable software and continuous digitisation all put our personal information at risk.
But the good news is that organisations can use those same AI tools and systems to fight this new wave of intelligent cybercrime. That’s why those with industry experience – like Peter Cooper, Broadcom Commercial & Enterprise Markets and Managed Partners Europe leader from February 2021 to April 2025 – strongly advise working them into existing cybersecurity systems to create a new line of defence.
Some of the most highly recommended implementations include:
- Upskilling talent on AI
- Governance structures to protect models from manipulation
- More secure asset inventories
- Internal defence systems that reduce reliance on third parties
The financial sector is particularly high-risk when it comes to cybercrime, rendering such changes practically non-negotiable. AI makes it possible to detect subtle anomalies that already exist in systems or anticipate potential threats before they even happen. Things that would have previously gone unnoticed until they become serious issues can now be destabilised in their infancy, strengthening and accelerating responses.
Risk management must remain at the heart of all new AI adoptions. Systems must be fit for purpose, supported by the most relevant insights and closely observed for any liabilities. Anything implemented by a third-party should be carefully vetted for malware or other threats, data should be thoroughly combed through, and frameworks must be trained to respond in the first instances of fraud or virus detection.
By 2027, it’s thought that universal access to AI for cybersecurity will become more likely, though efforts may be delayed by the fact that many systems are still underinformed or lack the resilience to protect large organisations against increasingly sophisticated threats. In some cases, AI models themselves may be the targets of cybercrime when it comes to dismantling digital defences.
That’s why the systems alone are not enough. They need knowledgeable, adaptable and accomplished people at the helm to ensure they advance properly, meet targets and have the ability to not only detect threats but also provide sufficient protection against them. Before we can fully enter the age of fully automated AI defence systems, upskilling human labour must be the priority. Paired together, human and machine intelligence spell a new era for cybersecurity.