Around 70% of journalists in Portugal have used generative Artificial Intelligence, AI, tools over the past six months to cope with the daily pressures of newsroom work, with one third reporting daily use. This finding is part of the report Artificial Intelligence and Journalism: Practices and Training in Portugal, produced by the Communication Observatory, OberCom, in collaboration with the Portuguese Journalist Training Centre, CENJOR, both members of the IBERIFIER consortium.
The research is based on an online survey distributed to active journalists in Portugal, which gathered more than 200 valid responses. The data indicate that AI adoption is now embedded in journalistic routines, with approximately 70% of respondents reporting usage, and 30% doing so on a daily basis. More than 70% also state that the technology has contributed to increased productivity in newsrooms, while only 3% report a decline in performance. Among the most widely used platforms are ChatGPT (66%), followed by Gemini (35%), Copilot (27%) and Perplexity (22%).
Ethical concerns
AI is primarily being used for operational tasks such as information research (80%), translation (58%) and transcription (53%). Its use is largely self-taught, with 71% of journalists stating that they learned to use these tools independently, while acknowledging limitations in crafting precise and effective prompts.
Despite widespread adoption, the report identifies a significant set of concerns. The findings point to a clear tension between the growing use of AI and uncertainty regarding its ethical and deontological implications. According to the researchers, this ambivalence may reflect a broader “existential questioning” within the profession, linked not only to the future of journalism but also to wider transformations in the labour market.
In terms of training, journalists consider current provision to be insufficient, sporadic and poorly structured. The report’s authors argue for a more integrated approach, combining technical skills with a strong ethical and deontological framework. Another notable finding relates to the lack of institutional support: 55% of respondents report that their organisations do not provide paid access to AI tools.
The report was produced by a team of researchers including Pedro Caldeira Pais, Miguel Crespo, Paulo Couraceiro, Ana Pinto-Martinho, Miguel Paisana, António Vasconcelos, Gustavo Cardoso and Vania Baldi.
