March 2025 – AI for Media Integrity, Public Research, and Trust Infrastructure

March 2025 continued the trend of institutional deepening of AI regulation and its integration into media literacy, scientific infrastructure, and civic resilience. A growing emphasis was placed on transparency, language inclusion, and fact-based content generation — both in public services and the creative industries.

European Union: AI & Media Literacy Fund Announced

On March 4, 2025, the European Commission launched a €100 million Media Integrity & AI Literacy Fund. The fund will support journalists, schools, and NGOs in developing fact-checking tools, bias detection platforms, and AI-assisted newsrooms. Details available via the Digital Europe Programme.

Italy: Launch of TRUST-LLM Public Research Facility

On March 6, the National Research Council (CNR) launched the TRUST-LLM lab in Pisa — Italy’s first public testbed for training and evaluating large language models. The lab’s focus includes dataset transparency, multilingual testing (including Italian dialects), and energy-efficient model design.

Malta: Public AI Dataset Registry Goes Live

Malta launched its National AI Dataset Registry on March 9, offering open access to datasets used in government AI systems (e.g. traffic, housing, immigration). The registry is searchable and includes metadata on collection methods and bias mitigation — an EU first. Hosted by MITA.

Spain: Deepfake Detection Hub Opens in Barcelona

On March 12, Spain opened the Deepfake Detection & Ethics Hub, co-funded by the EU and based at the University of Barcelona. The center supports newsrooms, law enforcement, and creators in identifying manipulated content. Its tools are made available via a public API.

Portugal: AI for Public Procurement Oversight

Portugal launched SIBILA, a GPT-powered assistant that reviews public procurement tenders and flags suspicious language or incomplete clauses. On March 16, the Court of Auditors confirmed that the system had already flagged over 200 cases for review.

Greece: National Call for AI in Multilingual Heritage

On March 19, Greece’s Ministry of Culture and Digital Governance announced a national funding call for AI projects that preserve linguistic heritage. Projects must use AI to document dialects, revive ancient Greek, or create museum companions in multiple languages.

Turkey: Launch of Anatolia NLP Benchmark

On March 21, Turkey’s AI Institute (TAI) published the Anatolia NLP Benchmark — the first comprehensive evaluation suite for Turkish, Kurdish, Arabic, and minority languages spoken across Anatolia. The benchmark will help fine-tune models on underrepresented languages.

OpenAI: GPT-4o Receives Multilingual Logic Boost

On March 25, OpenAI released a major logic and translation update to GPT-4o, improving reasoning across non-English queries. Tests show better performance in Italian, Greek, Turkish, and Spanish — particularly in mathematics, legal logic, and summarization.

🧠 Trending AI Tools – March 2025

AI tools leading conversations and adoption in Europe this month:

[SIBILA (Portugal)] (Government internal tool) – Procurement analysis assistant using GPT architecture.
GPT-4o Multilingual Update – Enhanced logic and cross-lingual reasoning now live in all tiers.
Gamma.app – Used at EU-funded labs and journalism conferences.
DeepL Write – Trusted by legal translators and cultural NGOs.
Leonardo AI – Generating visualizations for dialect preservation and museum guides.
Claude AI – Used in regulatory agencies and fact-checking teams.
Runway ML – Leveraged in anti-deepfake research and political campaigns.
Humata AI – Employed for digesting multilingual academic papers and policy briefs.
Tavily – Enhanced with EU regulation tracking by topic and date.
Perplexity AI – Used in AI literacy workshops and student research hubs.