The September 2023 installment of Crowdstrike’s Adversary Universe Podcast covered Artificial Intelligence’s potential for strengthening adversarial attacks. Adam Meyers and Cristian Rodriguez raised several key points to for the cybersecurity profession to consider:
- Interest of Threat Actors: Threat actors are actively discussing obtaining credentials and access to AI models like Chat GPT, indicating a growing awareness of AI’s potential advantages in malicious activities.
- Versatility of AI: AI can be employed for various purposes, both by adversaries and defenders. It ranges from automating exploit development and crafting complex malware to manipulating social engineering and deep fake content.
- Data Theft and Exploitation: AI facilitates more efficient processing and exploitation of stolen data, making it a valuable tool for both nation-state actors and criminals. This includes exploiting sensitive information, targeting individuals, and identifying system vulnerabilities.
- Vulnerability Discovery: AI accelerates the identification of software vulnerabilities by analyzing crash dumps and instrumenting fuzzers. This could lead to a higher rate of attacks and potentially more advanced exploit techniques.
- Adversarial AI: Adversarial AI involves manipulating AI models or algorithms to deceive or disrupt their functionality, creating a constant cat-and-mouse game between defenders and attackers.
- Privacy and Ethical Concerns: AI’s use in surveillance and recognition technologies raises concerns about privacy. Detecting and circumventing AI-driven detection methods is expected to become more prevalent.
- AI in Cybersecurity: Enterprises must carefully consider the use of AI in cybersecurity, ensuring its safe and effective implementation. Policies and education are essential to mitigate AI-related risks and vulnerabilities.
- AI Versus AI Warfare: As AI technology advances, it is increasingly likely that we will witness AI versus AI warfare in cybersecurity, where AI models are used to detect, manipulate, and counteract each other’s actions.
Meyers and Rodriguez underscore the complex and evolving nature of AI’s role in attack campaigns. As is the case with most technological innovations, the defenders can take advantage of AI to protect and defend assets. Ongoing research, education, and policy development are crucial to address these emerging challenges effectively.
In my opinion, it’s not whether cybersecurity professionals will lose their jobs to AI. It’s whether we will be replaced by someone who can use AI effectively.
Walt Houser 