The Fundación Maldita.es, with the support of the Patrick J. McGovern Foundation, will develop its own artificial intelligence infrastructure to improve and streamline fact-checking processes and respond to an increasingly fast-moving and complex disinformation ecosystem.
The project will enable Maldita.es to move towards the use of open-source models hosted on its own infrastructure, reducing reliance on commercial providers while making the use of artificial intelligence more sustainable, transparent and secure. It will also make it possible to adapt these models to specific fact-checking tasks, such as detecting hoaxes, disinformation narratives and potentially manipulated content.
Maldita.es has been developing technological tools to support the work of journalists and fact-checkers for years. These include chatbots that allow citizens to submit potentially misleading or disinformative content, systems that help sort and prioritise these submissions for editorial management, and La Buloteca, an open platform that allows users to follow the fact-checking process and encourages citizen participation.
Through Botalite, a technology company fully owned by the Foundation, this technology is already used by 23 organisations in 20 countries and four continents, in 48 languages. The improvements developed through this project will be made available across the network, allowing organisations to benefit from these advances without having to maintain their own infrastructure or engineering teams. The work will focus on three main areas:
Artificial intelligence in the service of fact-checking
New AI layers will be incorporated into existing systems to help detect scams and phishing attempts, identify suspicious content, automate logging and tagging functions, and support the work of fact-checkers. These tools will be developed in collaboration with editorial teams, with human oversight and testing against historical data before implementation.
Own and sustainable infrastructure
Maldita.es will move towards deploying open-source language models on its own GPUs, helping reduce costs and reliance on commercial APIs while also allowing the models to be specialised for specific tasks within the fact-checking process. This infrastructure will also serve the organisations already using Maldita.es technology.
Supporting organisations in high-risk contexts
The technological improvements will be incorporated into the infrastructure used by media outlets and organisations across the network, alongside training and support to facilitate their adoption. In addition, Maldita.es will continue to provide free access to its technology for independent media outlets in Latin America and strengthen its support for newsrooms operating in particularly vulnerable contexts.
Through this project, Maldita.es aims to strengthen the capacity of media outlets and public-interest organisations to investigate and combat disinformation, placing its own sustainable, open-source artificial intelligence infrastructure at the service of fact-checking.
How much funding will Maldita.es receive to carry out the project?
The project has been selected to receive 600.000 euros in funding from Patrick J. McGovern Foundation to be implemented between 1 August 2026 and 31 July 2027.
The Patrick J. McGovern Foundation (PJMF) is a philanthropic organization dedicated to advancing artificial intelligence and data science solutions to create a thriving, equitable, and sustainable future for all. PJMF works in partnership with public, private and social institutions to drive progress on our most pressing challenges, including digital health, climate change, broad digital access, and data maturity in the social sector.