Page 262 - Cyber Defense eMagazine RSAC Special Edition 2025

• Cloud-Native Era (2010s–2020s): The rise of cloud computing accelerated time to market and
operational agility, while bringing API attacks, supply chain threats, and mobile fraud.
• AI-Native Era (2020s–Present): AI is booming, delivering the biggest impact yet. AI is
revolutionizing decision-making, automation, and hyper-personalization. However, it is also
fueling AI-powered fraud, deepfake scams, and automated cyberattacks.

Despite these shifts, cybersecurity remains stuck in the cloud-native and software era. Cyber functions
have not yet declared 2025 to become AI-native – but they should.

AI-Native Threats are Real & Growing

AI introduces threats more sophisticated than ever. Some of the most pressing concerns include:

• AI-Powered Cyberattacks: Attackers use AI to automate social engineering scams, generate
malware that adapts, and even mimic user behaviors for phishing campaigns.
• Deepfake and Synthetic Fraud: AI-generated deepfakes manipulate voice, video, and text to
deceive facial recognition, trick executives, and spread disinformation.
• Autonomous Hacking Tools: AI-driven attack engines test for vulnerabilities, exploit them, and
adapt faster than human attackers.
• AI-Made Bias and Manipulation: AI manipulates financial markets, spreads political
propaganda, and interferes with decision-making through fake data.

Attackers have embraced AI as their weapon of choice. Defenders must do the same. AI-driven threats
magnify existing risks, increasing attack speed, variety, and impact.

CISO’s cannot afford to treat AI as an “LLM” or “data leakage” problem. To counteract AI-powered threats,
cyber functions must embed AI into their core operations, using AI to code, build, measure, and mitigate
threats in real time. They must embrace AI-native platforms to stay ahead of rapidly evolving threats.

The Big Difference with AI Threats

Unlike previous digital evolutions, AI is not just another attack surface—it is now the attack platform itself.
In other words, AI is the source of attack more than it is the target. Secondly, cyber functions remain
understaffed and rely on manual work and external teams to deliver defense. This dependency conflicts
with the business teams that are racing to embrace AI.

Starting in 2025, cyber needs an AI-native strategy that includes:

• Hyper Automation of Cyber Delivery: Cyber teams rely on IT, DevOps, and external security
vendors for implementation and enforcement – creating delays. AI-native solutions enable hyper
automation, faster deployment and better control. They also extend security across mobile, VR
and other digital spaces.
• Pre-Emptive Threat Management: AI already enhances fraud detection, but its true potential
lies in preventing threats before they emerge. Manual defenses can’t keep up with AI-driven

262