While it can be seen as a risk – an IBM study carried out in 2023 indicated that almost half of the executives surveyed feared that the adoption of generative AI would lead to new security pitfalls – it is also a prospect for perfecting security tools and automating some tasks.
So, in practice, what are the potential uses of AI in the context of cybersecurity? Let’s take a look.
Offensive AI: a constantly evolving threat
- Help in generating malicious code or scripts
- Help with phishing messages, voice cloning or deepfakes
- Help with processing stolen data
Defensive AI: anticipation and learning to stay ahead of the game
- Intrusion test automation
- Reverse engineering assistance
- Automated classification and prioritization of security events
- Analysis of malicious files and behavior
- Automated threat response
- Continuous learning to adapt and reinforce security
AI brings productivity gains at all levels, and for the moment it helps defenders more than attackers.
Today’s widely spread vast generalist models such as ChatGPT are costly and lack the precision needed to perform crucial cybersecurity tasks.
Also, models are specializing, becoming smaller and more efficient for specific tasks… and fine-tuning will continue to develop for applications in cybersecurity.
Want to know how we concretely use Artificial Intelligence
at HarfangLab? Have a look right here:
Deploying or changing EDR: a team effort
Who to involve in a project to deploy or change EDR, before and during the implementation of a cyber project.
Cybersecurity: choosing an EDR, key points to assess
How do you choose the right solution for your organization to protect your endpoints against threats? What criteria should you…
Why should organizations care about TDIR?
TDIR (Threat Detection Investigation Response) is a new approach to enhancing the analysis and action capabilities of SOC Managers.
Cybersecurity: how to optimize false positives
Reacting quickly to security incidents is crucial, but you also need the right information at the right time, with the…