Can AI Defend Us from AI?
AI is changing everything about how we build software. It speeds up development, helps teams manage complexity, and powers systems that learn and adapt in real time. But in our rush to innovate, we’ve introduced a new challenge.
- What happens when the systems we create become too complex to understand?
- And what if the only way to secure them is by using more AI?
This is the reality we face in 2025. Welcome to the age of the black box.
Modern digital systems are no longer built from the ground up. Instead, they are assembled from layers of open-source libraries, commercial tools, third-party integrations, and AI components. These tools are fast, modular, and powerful. They’re also often opaque.
Developers rarely know exactly how these black-box components work beneath the surface. And that’s a risk. Because when you can’t see how something functions, you can’t fully test it. You can’t prove it’s secure. You can’t guarantee it will behave as expected under pressure.
As more AI is integrated into critical systems such as hospital networks, energy grids, and autonomous vehicles, the stakes are getting higher. Traditional cybersecurity measures aren’t enough. It’s time for a new approach.
SECASSURED is a Horizon Europe research and innovation project that responds to this challenge. It brings together 19 leading research centers, companies, and public institutions from 10 countries, including Norway, Romania, Finland, Germany, Spain, and Greece. The project aims to create a future in which digital systems are trustworthy, certifiable, and secure from the inside out.
This isn’t a standard audit, but a comprehensive model for ongoing security assurance. Our approach spans the entire IoT ecosystem, from edge to cloud, and utilizes AI for real-time monitoring, testing, and validation. We use digital twins to imitate system behavior in a secure virtual environment, while our AI engines actively search for vulnerabilities and inconsistencies without interrupting development. We conduct tests alongside live operations to catch any threats before they can impact production. The best part? This continuous process ensures your systems stay safe and secure every day!
At ASSIST Software, our contribution to SECASSURED centers on building and operating the continuous integration environments that make this vision real. We design the infrastructure that supports ongoing testing, validation, and secure deployment of the AI-based tools developed within the project.
This means that every update, every patch, and every new feature is evaluated against security benchmarks before it goes live. It also means that development teams can continue to innovate quickly, without compromising on safety or compliance.
By transforming standard CI processes into advanced DevSecOps workflows, we are helping turn good intentions into guaranteed outcomes. In a world where change is constant, security must be continuous.
The work we’re doing with SECASSURED has real implications across industries. If you’re building infrastructure for telecom networks, delivering critical software for hospitals, managing platforms in the energy or mobility space, or integrating cloud-native systems for enterprise clients.
This project gives developers the tools they need to build with confidence. It gives security professionals the transparency they require to validate systems. It also gives organizations a framework for compliance that doesn’t slow them down.
For companies that have always viewed cybersecurity as a bottleneck, SECASSURED offers something new: security as an enabler. With AI handling the complexity, teams can focus on creativity, scale, and performance.
The old model of cybersecurity was based on isolated checkpoints. Test it once. Certify it. Ship it. Hope it holds.
The new model is dynamic. It verifies in real time. It adapts to new threats. It evolves with every change to the system. It uses AI to detect problems, predict them, and prevent them.
At ASSIST Software, we believe this is the only way forward.
You can explore more about the SECASSURED project and our role in shaping its future at assist-software.net/projects. The foundation of the approach lies in using AI to monitor, test, and validate systems in real time.
See Project Consortium: https://assist-software.net/project/secassured
See More EU Projects: https://assist-software.net/funded-projects/european-union-projects-partnerships
ASSIST Software is among the first companies in Europe to earn ISO/IEC 42001:2023 certification for Artificial Intelligence Management Systems. For clients, this means every AI initiative built with ASSIST Software is backed by a governance framework that addresses risk management, transparency, data quality, and long-term maintainability, the foundations that responsible AI adoption requires.
