Meet Amália: A New National AI
On July 1, 2026, Portugal unveiled Amália, its first national large language model (LLM). Named in honor of the iconic fado singer Amália Rodrigues, the AI also has a technical acronym: Automatic Multimodal Language Assistant with Artificial Intelligence.
Developed by a consortium of Portuguese universities and research institutions with government support, Amália is specifically designed for European Portuguese. Unlike closed, commercial models like ChatGPT, Amália is completely open-source. This means its model, datasets, and source code are freely available for businesses, researchers, and public institutions to use, modify, and build upon. The project is part of a broader European push for what's known as "sovereign AI"—the ability for nations to control their own digital infrastructure and reduce dependency on foreign technology.
Why Open Source Is a Game-Changer
The decision to make Amália an open-source platform is its most defining feature. While proprietary models from major tech companies are powerful, they function as 'black boxes.' Users can access them via an API but cannot inspect their inner workings, audit their training data, or modify their core architecture. Open-source models, by contrast, offer complete transparency. This allows developers to fine-tune the model for specific tasks, run it on their own secure hardware, and ensure compliance with local data privacy laws like GDPR. For Portugal, this means businesses can build custom tools without sending sensitive data to servers in other countries. It also fosters a local ecosystem of innovation, empowering smaller companies and public sector bodies to experiment with AI without prohibitive licensing fees.
Made for Portugal, By Portugal
Most mainstream AI models are trained predominantly on English-language data, which can lead to cultural biases and a lack of nuance when applied to other languages. Amália was built to solve this problem for European Portuguese. It was constructed upon a European foundation model, EuroLLM-9B, and then significantly enhanced by a team of over 60 researchers with vast datasets of European Portuguese text and images. This ensures the model understands the specific vocabulary, grammar, and cultural context of Portugal, a level of detail that larger, more generalized models often miss. This linguistic and cultural alignment is crucial for creating effective and trustworthy AI applications for the local population.
Powering a New Generation of Custom Tools
Amália was not created to be another consumer-facing chatbot. Instead, it’s designed as a foundational layer that organizations can use to build their own specialized applications. The possibilities are vast and tailored to Portugal's specific needs. Planned initial deployments will target the public sector, with applications including an AI-powered teaching assistant to help with lesson planning, a virtual guide for the country's museums and monuments, and a digital assistant for citizen services. The Portuguese Navy is also exploring its use for decision-support tools. For private businesses, Amália offers a platform to create everything from highly accurate customer service bots that understand local dialects to internal tools for analyzing market data with a deep understanding of the Portuguese context.
A Declaration of Technological Sovereignty
The launch of Amália is more than a technological milestone; it's a political statement. At the launch event, Prime Minister Luís Montenegro stated, "Europe's strategic autonomy is today, perhaps more than ever, tied to AI." This sentiment reflects a growing movement across Europe, with countries like France and Germany also backing their own AI initiatives. The goal is not necessarily to outperform the American tech giants head-on, but to ensure that Europe has control over the technology that will shape its future. By investing in sovereign AI, Portugal is safeguarding its digital future, fostering local talent, and ensuring that the economic and social benefits of artificial intelligence are retained within its borders.
