Equifax's Strategic AI Approach to Combat Cybersecurity Threats

Equifax is actively navigating the cybersecurity challenges that have escalated significantly since its infamous 2017 data breach, which cost the company $1.4 billion in remediation expenses. With artificial intelligence rapidly becoming a tool for cybercriminals, Equifax's leadership, particularly EVP and CISO Jeremy Koppen, is focusing on enhancing the company's defenses to combat these evolving threats.

Koppen, who joined Equifax in May 2025 after over a decade with Mandiant, highlights the significant uptick in automated attack patterns. “We’ve seen about a 30% increase in external attacks due to automation,” he notes. This surge, combined with a shrinking window for patching vulnerabilities, has created a pressing need for proactive cybersecurity measures.

Elevating Cybersecurity Hygiene

In response to the AI-related risks and existing security challenges, Equifax is doubling down on fundamental cybersecurity hygiene alongside the intelligent application of AI tools. A crucial part of their strategy includes expanding its passwordless authentication system to cover business partners, reducing the risk of social engineering attacks.

Equifax has implemented a comprehensive risk assessment tool that quantitatively analyzes business exposure. This tool enables the company to prioritize vulnerabilities effectively based on the potential impact of threats, allowing for a more measured response to emerging risks.

“Our focus is on ensuring that we can prioritize risks and manage them efficiently,” Koppen explains. Understanding which assets are most vulnerable, especially those exposed externally, is central to this effort. Multiple layers of defensive controls are being put in place to bolster security.

Leveraging AI for Incident Response

With security operations receiving upwards of 19.8 million alerts daily, Equifax has adopted an AI-driven approach to streamline incident response. Currently, around 50% of tasks in its security operations center are being handled automatically. This automation allows human analysts to concentrate on critical findings while AI offers context and knowledge to expedite the response process.

“AI is a tool for remediation,” Koppen states, emphasizing the importance of maintaining human oversight. Automation is also addressing past vulnerabilities, such as renewing and managing TLS certificates—an issue that contributed to previous breaches. With recent security reports underlining these advancements, the company is now better equipped to handle its certificate management automatically.

In the realm of software development security, AI is playing a transformative role as well. By integrating AI into the design phase, Equifax can conduct early code reviews, thus reducing the time needed for security validations from 46 days down to just 18 days. “This efficiency ensures that we maintain guardrails while verifying our outputs with human checks,” Koppen explains. Overall, their security consult times have decreased by an impressive 61%, and their AI systems can now identify over 213,000 vulnerability findings annually without impeding service delivery.

Proactively Addressing Rogue AI

While AI holds promise for identifying vulnerabilities, it also poses risks if not effectively controlled. “We need to ensure that our own AI can identify weaknesses within our environment prior to public disclosure,” Koppen emphasizes. He warns that AI agents capable of discerning vulnerabilities in a company's infrastructure may inadvertently detect inadequacies in their own security measures.

To mitigate these risks, Equifax employs strict identity-based access controls on its AI agents. “Locking down what an agent can access is vital. This approach is as important today as it was two decades ago,” he points out. The organization also spans protections that prevent data exfiltration and unauthorized use of infrastructure for third-party attacks.

Today, policy-as-code frameworks have replaced traditional manual checks, ensuring that every new AI agent goes through automated testing before deployment. Additionally, human review is mandated for critical applications, supplemented by ongoing monitoring to halt any erratic behavior exhibited by AI models. This proactive monitoring incorporates automated kill switches that enable instant rollback if necessary.

Koppen has been closely watching developments concerning AI agents escaping safeguards, emphasizing the importance of community cooperation in addressing these vulnerabilities. “It’s refreshing to see how engaged the community is in sharing ideas and solutions,” he remarks.

As Equifax continues to refine its approach to cybersecurity through the integration of AI across its operations, the company is striving to strike a balance between leveraging advanced technologies and maintaining stringent security protocols. By prioritizing foundational cybersecurity practices and embracing automation, Equifax aims not only to protect its data but to build resilience against future threats.