The Why AI team from the Lithuanian technology agency, led by artificial intelligence (AI) trainer Antanas Bernatonis, is introducing The Safe AI, an easy-to-implement data anonymization tool.
This tool addresses a pressing challenge: preventing data leaks when using artificial intelligence applications. The Safe AI automatically and locally detects confidential data in the browser, anonymizes it, and therefore prevents sensitive information from being transmitted to popular AI chatbots such as ChatGPT, Gemini, Claude, DeepSeek, or Copilot.
In this way, it tackles one of the biggest modern cybersecurity problems, where private data belonging to companies or public institutions is leaked through AI tools. As many as 38 percent of employees who use AI share sensitive work-related information with AI systems, often without the employer – and sometimes even the employees themselves – being aware of it.
Considering the growing importance of cybersecurity for organizations, this solution is particularly relevant. According to The Safe AI founder Antanas Bernatonis, it is essential not to fall behind current trends and to use AI as actively as possible, yet at the same time ensure the security of data belonging to individuals, legal entities, and public institutions.
A Simple Yet Effective Solution
The Safe AI is very easy to start using, as it works as a plug-and-play browser extension that can be installed in less than a minute. The tool relies on advanced pattern recognition algorithms that identify sensitive information in real time by recognizing recurring structures, such as customer details, document links, or source code.
Company administrators can also access key statistics and detailed reports on AI usage within the organization. This makes it possible to effectively manage and reduce the risks associated with unofficial AI use through non-corporate accounts, also known as shadow AI, when employees use unauthorized AI tools on their computers, increasing the risk of GDPR violations.
The data leak prevention functionality for AI tools will also help companies comply with strict data protection requirements, including the GDPR and the forthcoming EU Artificial Intelligence Act (EU AI Act), which foresees potential fines of up to 38,000,000 USD, 35,000,000 EUR and approximately 3,024,000,000 INR for non-compliance. This is also emphasized by Aurelija Rutkauskaitė, Partner and Head of Data Protection at the law firm Trinity Jurex.
By automatically anonymizing personal data and removing confidential information before it is submitted to LLM systems, the tool helps ensure compliance with GDPR, confidentiality, and cybersecurity requirements. At the same time, such solutions are becoming crucial for the practical implementation of the principles of the EU Artificial Intelligence Act, particularly in the areas of data governance, transparency, and risk control.
Importantly, data anonymization occurs locally. This means that the detection and removal of all sensitive data is performed on the user’s device, and confidential information does not leave the browser or reach any third parties or The Safe AI servers. This ensures that companies can use the most popular AI tools with text-based prompts more safely, without limiting employee productivity.
A Lithuanian Tool With International Potential
The project is being developed by the Why AI team, well known in Lithuania for its expertise in AI training and consulting, while founder Antanas Bernatonis is one of the leading figures shaping the practical adoption of AI in organizations. The expectation is that The Safe AI will contribute to safer use of AI tools in companies.
During training sessions, the team observed significant security gaps in how AI tools are used. For this reason, they are launching the tool for business clients and companies, while a simpler version will be offered free of charge to individual users so that feedback can be used to further improve the solution.
