Artificial Intelligence & Machine Learning,
Geo-Specific,
Next-Generation Technologies & Secure Development
Commission VP Henna Virkkunen Pledges Action in Tuesday Parliamentary Session
The European Commission has stepped forward to defend its strategies in addressing the rise of advanced artificial intelligence (AI) models that possess sophisticated capabilities for cybersecurity vulnerability assessments. This initiative comes amid rising concerns about a potential wave of AI-driven attacks expected to threaten the stability and security of the European Union (EU).
See Also: 5 Myths of AI & Machine Learning Debunked
This discourse gained momentum as, just two weeks prior, a collective of lawmakers heightened their calls for immediate action. Among their requests was gaining access to the Mythos model developed by Anthropic, recognized for its hacking capabilities. In a parliamentary session held on Tuesday, Commission Vice President Henna Virkkunen emphasized that the EU has already prepared itself with adequate tools to address the unanticipated challenges posed by these sophisticated AI models.
During her address, Virkkunen announced that the Commission is considering activating the EU Cybersecurity Reserve, which comprises a network of cybersecurity service providers that can be mobilized in times of crisis. She articulated, “We have not waited for these models to appear to anticipate future cybersecurity capabilities.” Alongside this, she committed to presenting a detailed action plan in the coming weeks that will integrate the expertise of EU stakeholders in both AI and cybersecurity to enhance readiness in facing upcoming AI threats.
Furthermore, the EU’s top technology official stated that the forthcoming Tech Sovereignty Package, anticipated to be introduced later this month, would contain provisions for “ensuring the maintenance, security, and integrity of our open-source digital infrastructures.” This initiative responds to growing cybersecurity risks attributed to the increased weaponization of AI.
Virkkunen also urged the Parliament to expedite the approval of amendments to the Cybersecurity Act. She argued that these revisions would empower the European Union Agency for Cybersecurity (ENISA) and the Commission with the necessary resources and authority to adapt to the evolving threat landscape driven by AI.
Additional calls were made concerning the prompt transposition of the Network and Information Security 2 Directive into national legislation for EU member states. This directive is crucial as it mandates cybersecurity responsibilities on critical infrastructure providers to bolster the entirety of the EU’s cybersecurity framework.
In her remarks concerning European access to the Mythos model, Virkkunen recognized the essential nature of these advanced AI models in safeguarding vital infrastructure amid an increasingly tense geopolitical landscape. However, she refrained from providing updates on the ongoing reluctance of Anthropic to allow access to its model. Instead, she emphasized the urgent need for public and private sectors within Europe to adopt “advanced cyber tools” that are already available to strengthen their defenses against potential threats.
“There is no reason to wait, and this is crucial for preparing for the upcoming risks,” stated Virkkunen, imparting a sense of urgency regarding the implementation of proposed measures.
In a recent development, OpenAI, a known rival of Anthropic, has granted select European authorities—including the Commission—access to its GPT-5.5-Cyber model, which bears similar capabilities to Mythos. This move illustrates the competitive landscape among AI developers in the realm of cybersecurity.
Additionally, France’s Mistral, currently the sole European entity involved in frontier AI, is reportedly developing its own hacking-capable model. This initiative aims to provide vital tools to major European banks amidst growing anxieties regarding security vulnerabilities. The banking sector in Europe is currently experiencing heightened alarm due to the looming threats posed by sophisticated AI.
In urging the EU to cultivate its own cybersecurity landscape, Virkkunen expressed her desire for Europe to minimize its reliance on external cyber-defense tools. She emphasized the necessity for creating the right conditions to foster sustainable cybersecurity business models capable of addressing present and future challenges. Nevertheless, no concrete proposals outlining how this goal should be achieved were presented.
Marilena Raouna, serving as a deputy minister from Cyprus, also addressed the parliamentary session, echoing concerns regarding Europe’s reliance on external entities for essential technological infrastructure. She pointed out that as Europe finds itself entangled in a highly competitive global atmosphere concerning advanced AI capabilities, it anticipates the introduction of the Tech Sovereignty Package as part of a broader strategic approach to autonomy.
In response to the discussions, several Members of Parliament urged more immediate action. Their suggestions included implementing mandatory red-teaming protocols, establishing an AI cyber tools certification scheme, and possibly activating the EU’s integrated political crisis response mechanism—a tool typically utilized for coordinating responses to conflicts, including the situation in Ukraine. However, Commission representatives did not provide feedback on these proposals during the session.