FIS has expanded its partnership with Anthropic by joining Project Glasswing, an initiative that uses advanced artificial intelligence models to identify and address potential vulnerabilities in software supporting critical infrastructure.
Through the initiative, FIS is testing Mythos 5, Anthropic’s most advanced AI model, as an additional component of its internal cybersecurity program. The company is using the model to scan and evaluate its systems for security weaknesses that could affect financial institutions and businesses relying on its technology.
Project Glasswing is a controlled-access program for organizations that develop or maintain foundational software. Participants use Anthropic’s frontier AI models for defensive security work, including reviewing complex codebases and identifying vulnerabilities that may be difficult to detect through conventional methods.
FIS operates technology that supports core banking, payment processing and money movement for thousands of institutions worldwide. Because those systems are deeply connected to global financial infrastructure, software vulnerabilities could have significant consequences for banks, businesses and their customers.
The company said its participation reflects an effort to apply the same security standards to its own infrastructure that it expects from the products and services delivered to clients.
Mythos 5 will function as an additional layer within FIS’ broader security program rather than replacing existing cybersecurity teams, controls or testing processes. The technology can help analyze large amounts of code and surface potential risks for further investigation by security professionals.
Frontier AI models may improve defensive security by recognizing complicated patterns across software components, dependencies and configurations. These capabilities could help organizations identify weaknesses earlier and prioritize remediation before vulnerabilities are discovered or exploited by attackers.
Financial technology platforms are especially complex because they must operate continuously, process high transaction volumes and connect with numerous financial institutions, payment networks and third-party systems.
Changes to one component can affect other applications or infrastructure, making comprehensive security testing difficult at scale. AI-assisted analysis could give security teams another way to evaluate interconnected systems while maintaining human oversight over findings and remediation decisions.
FIS’ participation in Project Glasswing is separate from its commercial AI partnership with Anthropic. That relationship involves the use of Anthropic-powered AI agents and other capabilities within financial services products and workflows.
The Project Glasswing collaboration is instead focused on defensive cybersecurity and strengthening the software infrastructure operated by FIS.
Both initiatives reflect the company’s broader approach to introducing advanced AI into environments where reliability, security and trust are essential. Financial institutions generally require stronger governance and controls than organizations using AI for lower-risk applications.
FIS also shapes its cybersecurity strategy through participation in the Financial Services Information Sharing and Analysis Center and the Financial Services Sector Coordinating Council.
Those organizations help financial institutions, technology providers and government agencies share intelligence about cyber threats, vulnerabilities and emerging risks affecting the financial system.
FIS also works with regulators and other industry participants as it evaluates new technologies and security practices.
By joining Project Glasswing, FIS will contribute to a broader effort to determine how frontier AI models can be used responsibly to defend critical software. The initiative may also help participating organizations develop practices for testing AI-generated findings, managing false positives and ensuring that sensitive systems and data remain protected.
The partnership comes as financial services companies face increasingly sophisticated cyber threats and growing pressure to secure large, interconnected technology environments.
Attackers are also beginning to use AI to accelerate vulnerability discovery, phishing, malware development and other malicious activity. This shift is encouraging defenders to evaluate how advanced models can help security teams respond at comparable speed and scale.
FIS’ use of Mythos 5 is intended to strengthen proactive security by identifying potential weaknesses before they develop into operational or customer risks.
The company believes the initiative can support the resilience of its own systems while contributing to the broader security of the financial services sector.
