From lost corporate data to complete control
Would you like to determine if your company is exposed to risks regarding the new AI legislation? Our experts are ready to conduct a deep dive into your current processes for a comprehensive Risk Assessment. Our experts will gladly take a deep dive into your current processes for a thorough Risk Assessment.
Everyone is talking about AI, but what about the foundation: data? Whether your data is secure extends far beyond the database alone.
But what are the biggest risks, exactly?
AI is on everyone’s mind, but what about the data that forms its foundation? Data security involves much more than just the database. Read on below to learn about the major risks.
Most companies are unaware of the dangers surrounding US legislation such as FISA and the US Cloud Act. Under FISA regulations, agencies like the CIA and FBI do not require prior judicial approval to access the data of European citizens.
Given current geopolitical tensions, storing data with American providers is simply not a smart move. Although Big Tech promises to bypass this legislation through legal contracts and additional security layers, the law remains the law. Businesses urgently need to become aware of this.
Europe places a strong emphasis on protecting its citizens. Through the GDPR, the law requires organizations to handle personal data with care and transparency.
Additionally, the AI Act has recently been introduced, which further regulates and frames the use of artificial intelligence within the European Union. This regulation can cause tension with American legislation in certain cases. Since many technologies from large international players have not been specifically developed with European data sovereignty as a starting point, legal and operational challenges may arise.
For executives, this means that they remain legally responsible for the processing of data, even when the actual control over that data is limited.
An additional challenge is the so-called ‘black box’ problem. It is often not fully transparent how large language models (LLMs), developed by international technology companies, handle data and where it is exactly stored. In the past, there have been incidents where users gained access to the chat history of others, raising questions about data protection.
Additionally, the possibilities for completely deleting data are limited in some cases. When data is used for training models or supporting AI applications, it is not always clear how and where this data is further processed. This makes it difficult to guarantee complete control or traceability afterwards.
At the same time, general AI models do not always provide the specific context of an organization. As a result, they do not always optimally align with internal processes or strategic objectives, highlighting the need for more context-oriented and business-oriented AI solutions.
AI makes it very easy for malicious actors to hack.
In fact, AI is a scaling tool: if data is not properly protected and cybersecurity is not prioritized, AI puts a motor behind the weaknesses of an organization.
AI is one of the most powerful technologies ever developed. Just like with the atomic bomb, strict rules are needed before disasters happen, and AI is no different.
Big Tech is currently handling this recklessly; they sell LLMs to 'no-code' developers who simply lack the technical knowledge to intervene when an AI needs adjustment. By working directly with cloud models like those from OpenAI and Anthropic (Claude), a safe 'sandbox' (a protected testing environment) is often missing. This can become extremely dangerous.
However, Big Tech will not stop selling these subscriptions themselves, as they have every incentive to maximize profits.
An increasing number of cases are coming to light where citizens and employees fall victim to irresponsible AI usage lacking clear guidelines. There are already reports of rising suicidal tendencies among vulnerable users, while misguided startups blindly chase 'unicorn status'.
The cause? Big Tech has developed LLMs in such a way that they seem to be humanity's best friend. This creates a collective delusion in which people become almost hypnotized by the technology.
Regulating and enforcing accountability on such reckless AI is therefore a matter of extreme urgency. Engineers and experts are desperately needed to closely guard these boundaries and to detect and neutralize dangerous patterns early on.
Today, companies pay foreign tech giants through licenses and cloud subscriptions. This ecosystem creates a 'vendor lock-in' that sustains itself.
This creates a dependency on products from a single supplier, making it difficult, expensive, and time-consuming to switch. The longer a company waits, the smaller the chance becomes to invest in its own AI infrastructure and reclaim digital independence. Safe
The declaration of digital independence
Secure
Sovereign
EU
Why flanders stops funding Silicon Valley
From conflicting legislations to geopolitical tensions and trade wars: it is an absolute no-brainer for Europeans that we must reclaim our digital independence.
All of Europe is currently on hot coals. And here is Thor. With a team of top lawyers and engineers, we are working within data nerds on the future.
Thor
With DataNerds, you choose fully responsible AI without the technical headaches because we take care of everything from A to Z:
Hardware-as-a-Service: No high upfront investments in servers or data center facilities. You simply subscribe to your secure digital environment as a service.
Plug-and-play: Thor comes with a user-friendly interface, ready for use by your employees.
Fully managed: Our engineers handle integration, security, and updates in the background, fully aligned with European AI legislation.
This is how you transform AI from an unpredictable expense into a valuable, proprietary asset that protects your company’s competitive edge.
90%
As many as 90% of companies are currently experimenting with AI, without considering the legislation and the future of Europe.
How to start?
Before our engineers write a line of code or connect your systems, we work thoroughly with a strategic audit.
Audit
Together, we will evaluate your current AI systems and benchmark them against the new EU regulations. Next, our experts will lay the foundation for a secure and intelligent data infrastructure.
The result?
A concrete action plan that provides total clarity on how to leverage AI in a way that is highly effective, secure, and legally authorized.
Product
Our engineers build Thor.
You retain full control and easily operate Thor via your own dashboard or a custom interface.
Monitor
You choose:
A full transfer to your internal IT team (including a structured legal exit) or a continuous collaboration where our experts handle the ongoing training and maintenance of Thor.
Frequently asked questions (FAQ)
Here are some frequently asked questions about Thor and the IT infrastructure we recommend and implement.
This solution allows organizations to regain absolute sovereignty over their corporate data. By combining private hardware at LCL with the European Mistral AI engine, a digital fortress is established that remains 100% independent of Big Tech.
This offers the speed and flexibility of the cloud, but with the physical security of your own vault on Belgian soil.
The AI Act imposes strict requirements on the transparency, traceability, and safety of AI systems. Our engineers maintain oversight of the product they have developed. Thanks to feedback reports and Monitoring-as-a-Service (MaaS), we do what is necessary to ensure Thor operates under the strict European conditions.
The fundamental difference between Thor and ChatGPT lies not in how smart they talk, but in sovereignty, control, and architecture. You can compare it to the difference between a public bus and an armored company vehicle owned.
1. Legal and physical sovereignty (The fortress)
ChatGPT: is an American product (OpenAI). Even with their enterprise licenses, it remains a 'black box' that runs on American cloud infrastructure (Azure). Legally, as a European multinational, you are directly in the danger zone of the EU AI Act and GDPR legislation, not to mention being exposed to FISA legislation and others like the American Cloud Act.
Thor is 100% European and sovereign—from the LLM down to the very bare metal of the servers. Thor was explicitly developed based on European standards. What makes Thor even more compelling is that it runs entirely on secure EU databases and effectively speaks the language of your business.
2. No 'Black Box', but a transparent architecture
ChatGPT: There is no transparency regarding the model's training data, how its weights are fine-tuned, or what data will arbitrarily be used tomorrow to optimize the algorithm. Ultimately, you lack total control over the back-end.
Thor: We develop fully customized. The reason our audits are important is not only to understand the current systems of the company but especially to outline a clear prototype plan. This way, we create the ideal synergy between your data and AI in a safe, legal manner. Moreover, Thor is built so that every action and event is displayed transparently.
Thor: Is specifically trained by our data engineers on your years of accumulated data. Thor handles the automation within your enterprise and manages your business processes autonomously, without any risk of data leaks or external 'data poisoning'.
That can happen if it is a European system. In fact: that is exactly what the technology is built for. If there is already a well-functioning European ERP system running (like SAP, Odoo, or another platform that operates outside the American Cloud Act), there is absolutely no need to change anything.
Read more about integrations.
