Download this article in PDF format.
Cybersecurity breaches continue to escalate, fueled in part by more sophisticated attacks and the growing use of artificial intelligence (AI). Hackers can now create more convincing phishing emails, scan for weaknesses and test systems faster than they could 10 years ago, giving organizations less time to detect and respond to threats.
The Nitrogen Ransomware Group’s attack on Foxconn is one of the latest breaches to hit close to home for electronics buyers and supply chain managers. According to Cyber Magazine, the Taiwanese electronics manufacturing giant, whose customers include Apple, Google, NVIDIA and Sony, was targeted in a ransomware attack that allegedly involved eight terabytes of data spanning more than 11 million files.
The ransomware group claimed the stolen files include confidential customer data, projects and drawings tied to major technology companies. Foxconn has not confirmed those claims, but it did acknowledge that some of its North American factories suffered a cyberattack. According to the publication, the company’s “cybersecurity team activated its response process and took steps to maintain production and delivery as affected factories resumed normal operations.”
Cybercrime Gets More Expensive
The attack on Foxconn is one example of a cybersecurity problem that continues to grow in both scale and cost. Cybersecurity Ventures projected global cybercrime costs would reach $10.5 trillion annually by 2025, up from $3 trillion in 2015. As ransomware attacks, data theft and supply chain breaches continue to proliferate, IBM says the average breach cost is now $10.22 million for U.S. companies and $4.44 million globally.
AI is fanning the fire by helping cybercriminals automate phishing campaigns, create more convincing fake messages and identify vulnerabilities faster than they could just a few years ago. Panda Security says:
- Reported cybercriminal losses have doubled between 2022 and 2025, a growth spurred by AI.
- 99% of organizations have sensitive data dangerously exposed to unsanctioned AI tools.
- Among data breaches involving AI attacks (16% of breaches), 37% used phishing and 35% used deepfakes.
- 82.6% of cybersecurity professionals are concerned about the increasing sophistication of AI-powered phishing attacks.
“We’ve entered a new era of cybersecurity, as both defenders and attackers leverage artificial intelligence, creating what experts describe as an ‘AI arms race,’” Panda says. “Hackers can weaponize AI to generate highly personalized messages that mimic the writing style of trusted contacts or organizations.”
At the other end of the spectrum, AI is also helping companies defend themselves against cyberattacks and more proactively monitor, identify and thwart potential cybersecurity issues. For example, Panda says AI helps maintain 99% accuracy while also saving companies an average of $1.9 million in breach costs.
Using AI to Police AI
With AI agents able to autonomously perform reconnaissance, exploit vulnerabilities and move laterally across networks—and all without any human intervention—SentinelOne says companies are switching to agentic AI models that can handle everything from triage to 24/7 monitoring to automated incident response.
So the same tools that attackers are using can also help companies defend themselves. Instead of waiting for IT teams to chase down every alert manually, for example, agentic AI can help spot unusual activity, prioritize the most urgent issues and put basic controls in place faster.
SentinelOne says IT departments are also using agents to map AI usage across the organization and enforce data boundaries across endpoints. The goal is to find out whether employees are feeding sensitive data into public AI tools. The agents can “block unauthorized tools at the network level,” it says, “while offering approved, sandboxed versions that sanitize data before it leaves the corporate environment.”
