The AI industry has a security problem, and it’s not just digital. Rising opposition to artificial intelligence, ranging from physical threats against people and property to increasingly sophisticated cyberattacks, is forcing tech companies to rethink how they protect everything from their data centers to their employees.
Roughly 73% of global organizations have already integrated AI into their cybersecurity defenses. Nearly all of them plan to keep investing in AI-driven security solutions through 2025.
The core problem is what security researchers have started calling an “AI vs. AI arms race.” Adversarial machine learning techniques, where attackers use AI to probe and exploit other AI systems, are becoming standard practice. Data poisoning, where malicious actors corrupt the training data that AI systems rely on, has emerged as a particularly insidious vector.
State-sponsored actors have entered the chat too. North Korean operatives have reportedly used large language models from OpenAI, Google, and Anthropic for malicious activities, including targeting crypto projects.
The crypto sector is ground zero for AI-enabled fraud
