AI Use Cases in Journalism: An Open Inventory of Examples
Updated: Oct 7
Artificial intelligence is transforming the media, but understanding how it is doing so requires going beyond predictions, trends and announcements about new technologies. What problems are media organizations trying to solve with AI? In which processes are they incorporating it? What specific applications are they developing and what technologies are they using?
I have been studying the relationship between artificial intelligence and communication for nearly 15 years. In addition, since 2018, I have been collecting and analysing use cases of artificial intelligence in the media, initially as part of my doctoral research and subsequently as part of my research work on AI, communication and culture.
The result is a corpus that has been built progressively over the years. The cases were not initially collected with the aim of creating a public database, but as part of my own research and documentation work. Over time, this inventory grew and also became a tool for observing the evolution of the uses of artificial intelligence in the sector.
In 2025, I decided to share this inventory publicly, making a database of the cases collected available, together, whenever possible, with a link to the original source. The desire to share knowledge and facilitate access to sources has accompanied the project ever since.
Now this inventory is taking another step forward.
Explore AI use cases in the media
Explore the database of documented cases of artificial intelligence use in the media and filter cases by country, year, category, production phase, and objective:
The search tool therefore grew out of that first inventory shared in 2025. It is not an independent project, but rather the evolution of a corpus built over many years and progressively transformed into an open reference resource.
The tool allows users to explore the cases through their own questions: what uses are being made of AI, in which countries or media organizations, in which formats or with which technologies, among other possibilities.
The database does not aim to provide an exhaustive or definitive snapshot of AI adoption in the media. Its purpose is to make available a corpus of documented cases that can be used to explore the phenomenon, identify patterns, compare initiatives and formulate new questions.
Below I explain how this corpus has been built, what criteria have been used to identify and classify the cases, and what its limitations are.
Methodology
The systematic collection of AI use cases in the media began in 2018, within the framework of my doctoral research, and has continued ever since. Cases have been identified and analysed individually. The subsequent systematisation of the information has also been supported by artificial intelligence tools for structuring, standardising and classifying the corpus, always with human verification and review.
The research underpinning this project takes a multidisciplinary perspective. In addition to research on communication, journalism and technology, it incorporates contributions from fields such as philosophy, ethics, technoethics, technology and society studies, and other disciplines related to algorithmic mediation.
The academic references underpinning this research can be consulted extensively in my doctoral thesis and in the book Artificial Intelligence, Ethics and Communication: The Human Factor in the Age of Algorithmic Mediation (Editorial UOC, 2024).
What do we mean by artificial intelligence?
This project adopts the definition of an artificial intelligence system established in the European Artificial Intelligence Act (AI Act):
Machine-based system that is designed to operate with varying levels of autonomy and that may exhibit adaptiveness after deployment, and that, for explicit or implicit objectives, infers, from the input it receives, how to generate outputs such as predictions, content, recommendations or decisions that can influence physical or virtual environments.
This definition makes it possible to approach artificial intelligence from a broad and also historical perspective. The corpus therefore includes applications developed before the widespread adoption of today's generative AI when they use techniques that can be considered part of artificial intelligence, such as natural language generation (NLG), machine learning, natural language processing or speech synthesis.
At the same time, not all algorithmic automation is necessarily considered artificial intelligence. When a system is limited to executing rules, templates or predefined workflows, and the available sources do not allow the use of AI techniques to be identified, this distinction is maintained in the classification. The aim is both to avoid excluding historical applications of AI and to avoid attributing this status to systems whose technical nature is not documented.
What is a use case?
For a case to be included in the corpus, it must correspond to a specific application of artificial intelligence in the context of a media organisation or communication organisation.
We understand a use case as:
A practical and specific application of this technology intended to address a concrete need, improve a process, optimise resources or generate value, whether at an operational, strategic or service level.
This definition makes it possible to distinguish concrete use cases from mere statements of intent, generic announcements about artificial intelligence, or content that simply mentions the technology.
Each case is analysed on the basis of the information available about the initiative. Whenever possible, the database includes a link to the original source so that the case can be consulted directly.
Cases are identified through a review of diverse sources: media organisations, institutions, research projects, databases, specialised publications, podcasts and other resources related to the application of artificial intelligence in communication.
This same approach is present in other resources I have developed, such as the compilation of more than 70 guidelines and codes of practice on the use of artificial intelligence in journalism and the media and communication organisations worldwide, which I also continue to update.
All these projects share the same underlying logic: to facilitate structured, rigorous and useful knowledge for people who research, teach, work or make decisions in this field.
Classification and standardisation
The information in the cases is standardised using categories and a controlled vocabulary, with the aim of facilitating consultation and comparison.
When different sources describe the same initiative, an effort is made to identify it as a single case and avoid duplication.
Artificial intelligence tools are part of this systematisation process, particularly for tasks involving structuring, classification and standardisation. However, they do not replace the analysis or verification of the cases.
One of the project's principles is not to present as fact anything that the available sources do not allow us to establish.
The information published by media organisations is highly uneven. In some cases, it is possible to identify the technology used with precision; in others, the information is limited to describing the application, objective or outcome.
Human intervention is particularly difficult to determine in many cases. The fact that an organisation describes an application as automated does not necessarily tell us what degree of human supervision or intervention exists in practice.
For this reason, I do not classify the degree of human intervention when the available information does not allow it to be determined with sufficient confidence. I prefer not to assign a level rather than introduce an inference that could be incorrect.
The same criterion applies to technical architecture. A media organisation may communicate which tool, model or technology it uses without explaining how the system is built, what other components are involved, or how they are integrated.
Architecture is not inferred when it is not documented. Reconstructing it from partial information would introduce a degree of speculation that I do not consider methodologically justified.
This caution is particularly important in historical cases. The available documentation may describe an application as “automated” or “algorithmic” without specifying the technology used. In these cases, a specific technology is not attributed through inference. When the source explicitly identifies a technique, such as NLG, it is recorded even if it belongs to an earlier generation of AI systems.
Limitations
The database is not an exhaustive census of the use of artificial intelligence in the media. It contains cases identified and documented on the basis of the sources available.
There is a significant visibility bias: media organisations do not communicate all the projects they develop, and some are much more transparent than others. This difference may be related to the country, institutional culture, communication policies, available resources or an organisation's tradition of transparency.
There is also a bias resulting from the research process itself. The corpus has paid particular attention to media organisations that are pioneers in experimenting with artificial intelligence and to organisations that actively communicate their initiatives, as well as to media organisations in Catalonia and Spain, in line with the origins and objectives of the research.
For this reason, some particularly active and transparent media organisations, such as the BBC or The New York Times, as well as certain Spanish media organisations, have a prominent presence in the corpus.
This means that the distribution of cases cannot be directly interpreted as a measure of AI adoption. A country, media organisation or organisation with more documented cases may reflect greater availability of information or a greater willingness to communicate its initiatives, rather than a proportionally higher level of adoption.
Likewise, the absence of a case from the database does not imply that the practice does not exist.
The database is a living project and aims to incorporate many more cases as new initiatives are located, documented and analysed.
Cases can also be expanded, corrected or updated when new information becomes available. This is particularly relevant in a field such as artificial intelligence, where tools, models, applications and projects evolve rapidly.
For this reason, the database has been conceived as a resource for exploring, comparing and formulating questions, rather than as a source intended to produce a single interpretation of the sector.
Since both the explorer and the table are conceived as collective and evolving tools, if you know of a case that is not included and leave it in the comments or send it to this contact, I will be happy to add it, always citing the source correctly. And if you find this table useful for your practice, teaching or research, I would be pleased to hear from you and to have your citation (how to cite).
If you prefer to consult the table, you can do so by scrolling both horizontally and vertically with the cursor. Simply place the mouse over the table to enable scrolling. I have highlighted cases from Spanish media organisations so that they can be found more easily.
Sources
This resource is meant to be a collective, evolving tool. If you know of a relevant case that’s not yet included, I’d be happy to add it, always with proper attribution. And if you find this table useful for your work, teaching or research, I’d love to hear about it.
You can contact me at info@patriciaventura.me to suggest examples or ideas. The table will continue to grow with what we build together.
To carry out this work of collecting and analyzing use cases of artificial intelligence in the media, I have found several sources particularly helpful, which I share below. These include the database from the JournalismAI project at the London School of Economics, the AI Journalism Starter Pack, the database published by the Associated Press, the list of generative AI use cases in media by Kalle Pirhonen, the blog of the Master’s in Innovation in Journalism at the University of Elche, the journal Cuadernos de Periodistas by the APM, the Newsroom Robots podcast by Nikita Roy, the Generative AI in the Newsroom blog by Nicholas Diakopoulos, as well as other experts who openly share their knowledge. For some time now, I have also been committed to disseminating many of these resources through my website, where I keep an updated list of sources on use cases and tools related to artificial intelligence in journalism.
This initiative is part of a line of work I’ve been developing for over 15 years to understand how AI is transforming communication. I’ve approached this through dissemination and academic research—including a PhD on ethics, AI and communication defended in 2022—and, within that work, produced a report and a code of ethics for responsible AI adoption, published by the Catalan Information Council in 2021. But also through ongoing collaborations with media outlets, public institutions and training spaces, always with a critical and constructive perspective. Trying, whenever possible, to ensure that technology debates do not lose sight of the human dimension.
[Article written and translated with the assistance of a generative AI tool]
Explore AI use cases in the media
Explore the database of documented cases of artificial intelligence use in the media and filter cases by country, year, category, production phase, and objective:
[This article has been automatically translated with AI].
Comments