US company Anthropic is facing mounting pressure from the European Union over its decision not to make a new artificial intelligence model with extraordinary hacking capabilities available to European regulators. Tensions between Brussels and the company have escalated significantly in recent days, pointed out the server Politico.
The model, called Mythos, was introduced in early April and, according to the company itself, can outperform most people in detecting and exploiting software vulnerabilities. It was because of concerns about its potential misuse that Anthropic decided to severely restrict access to the technology. However, this has caused discontent among European institutions, who complain about the lack of information and testing capabilities.
The situation will be discussed in the European Parliament, where representatives of the European Commission and ENISA are to discuss the risks associated with this instrument. Anthropic will not attend the meeting, which Dutch MEP Kim van Sparrentak said is a serious concern. It was she who initiated the debate.
Brussels and individual member states are beginning to lose patience because the company only shares the model with a limited number of mainly US organisations, leaving European authorities with little room to prepare for the potential impact of the technology. European Commission spokesman Thomas Regnier said that from August 2026, when the AI Office becomes fully operational, the EU will force access to similar models if necessary.
The tension was further heightened by a letter from thirty MEPs to Commissioner Henna Virkkunen demanding the creation of a European plan to protect key systems from hacking risks. According to Dutch MEP Bart Groothuis, Europe is currently „not at the table“ and needs to apply much more pressure.
There are also concerns in the financial sector. European banks are trying to assess the risk posed by the spread of similarly advanced instruments. While US-based JPMorgan Chase is one of several organisations that have access to the model, European banks are not among them.
The contrast between the EU and the United States is stark. In the US, technology companies such as Microsoft, Google DeepMind and OpenAI are already working with the government to test their models for national security before they are launched.
European leaders are therefore calling for greater transparency and international cooperation. However, according to Greek Finance Minister Kyriakos Pierrakakis, the situation is complicated by the tense relations between the EU and the US. Nevertheless, he said Europe has no choice but to seek common ground, as the impact of AI will be substantial.
gnews.cz - GH
Czech 
English 
Spanish 
French 
Italian 
Russian 
Chinese 
German