Quantum Threat and Artificial Intelligence: How to Protect Yourself in the Digital Future

At first glance, quantum technologies and artificial intelligence (AI) seem to be the driving forces of progress. However, at the Synergy 2025 conference, experts explained that they are also becoming the main challenges in cybersecurity. The world is rapidly approaching a point where conventional defences will be useless and the main weakness is not technology, but ourselves.

Christophe Bianco of Thales Luxembourg said, "Our entire digital infrastructure is built on trust in cryptography. Tomorrow, that trust could disappear." The culprit is quantum computers, which can break most of today's encryption algorithms. Hackers are already collecting encrypted data now so they can decrypt it later - a strategy dubbed "harvest now, decrypt later".

Despite the fact that many consider the threat of quantum hacking to be 10 years away, China is already claiming breakthroughs in this field. Experts at the conference agree on one thing: Prepare now.

Jean-Francois Mairlot of IBM Luxembourg suggested a practical methodology:

• Risk assessment: scanning all systems and applications for vulnerable algorithms. IBM offers tools for compiling a "cryptographic bill of materials" (CBOM).
• Prioritisation: identify the most critical systems and focus efforts on them.
• Remediation: begin implementing algorithms that are resistant to quantum attacks, such as those selected by NIST (US National Institute of Standards). IBM is involved in the development of three of four such algorithms.

In early 2025, IBM also announced a partnership with Vodafone - they are introducing quantum-resistant security for smartphones.

Mark Tehrani of startup Cyberseq and Steven Maas of BeNeLux called for "crypto-flexibility" - the ability to quickly update security standards as new threats emerge.

AI is not only a defence tool, but also a target for attacks. Mairlot distinguishes two directions:

• "AI for Security": using AI to protect systems.
• "Security for AI": protecting the AI models themselves from attack and manipulation.

All leading security solutions now include AI. But that doesn't negate the threats posed by AI itself. Mairlot recommends exploring the OWASP Top 10, a list of top web application vulnerabilities that applies to AI systems as well.

The main danger, according to Christophe Bianco, is not technical, but human. AI has become too accessible, and most users don't understand how it works. This makes them vulnerable to fraud and manipulation. One of the most common types of fraud in Luxembourg today is the substitution of bank details in PDF invoices. And in such cases, it's not about deep forgery at all, but a loss of critical thinking skills.

The transition to a quantum-safe future takes time and resources, but, as experts emphasised, it is no longer possible to postpone. Similarly, with AI, it is important not only to admire its capabilities, but also to understand what new threats it poses.