Tomi Fischer tarkasteli maisterintutkielmassaan tekoälyn käyttöönottoa suomalaisissa mediataloissa haastattelemalla 18 ammattilaista toimitus-, johto- ja teknologiatehtävistä. Tutkimus paljasti niin sanotun tehokkuusparadoksin: vaikka tekoäly nopeuttaa yksittäisiä tehtäviä, se lisää samalla sisällön tarkistamisen tarvetta, vaatii aikaa järjestelmien hallintaan ja heikentää toimittajien syvällistä aineistoon perehtymistä.
Lisäksi tutkimus toi esiin organisaatiotason ristiriitoja – tekniset tiimit korostavat tehokkuutta ja käyttöastetta, kun taas toimittajat painottavat laatua ja autonomiaa. Suomea koskevat kieliongelmat tuovat lisähaastetta, sillä työkalut toimivat huomattavasti heikommin suomeksi kuin englanniksi, mikä pakottaa toimitukset tekemään päätöksiä teknisten rajoitteiden pohjalta.
Tutkimus suosittelee neljää periaatetta eettiseen tekoälyn käyttöön: ihmisen valvontaa, ympäristövastuuta, laadun asettamista tehokkuuden edelle sekä läpinäkyvyyttä yleisölle. Yhteenvetona se korostaa, että kysymys ei ole vain siitä, nopeuttaako tekoäly journalismia, vaan siitä, voiko se parantaa sitä säilyttäen journalismin demokraattisen tehtävän.
Scholarship Blog: The Efficiency Paradox – How AI affects journalist workflows and organizational dynamics
When journalists at Finnish media organizations first started using AI tools for transcription and translation, the promise seemed clear: faster workflows, more time for actual reporting, and greater efficiency across newsrooms. But new research from Aalto University reveals a more complex reality behind these technological promises.
That feeling when you use an AI tool to generate a headline and realize it completely missed the nuance of your story. Or when you spend more time learning how to prompt an AI system effectively than you would have spent writing the content yourself. These experiences, familiar to many Finnish journalists today, point to what researchers call an ”efficiency paradox” in AI implementation.
The comprehensive study, which examined AI adoption across Finnish media organizations through interviews with 18 professionals, found that while AI accelerates individual tasks, it simultaneously creates new cognitive burdens that can eliminate the promised productivity gains entirely. The research was conducted across editorial production, editorial leadership, and technical development roles, revealing differences in how these groups experience AI integration.
”The text mess comes quicker. There’s more and more text you can have, but then the checking part takes more time because you have to check every sentence,” explained one journalist interviewed for the study. This sentiment captures the core of the efficiency paradox: AI tools create verification requirements that traditional methods simply didn’t demand.
The paradox operates through three mechanisms that emerged from the research. First, journalists must now fact-check not just their sources, but their tools themselves. Second, mastering AI systems requires significant time investment that competes with deadline-driven editorial work. Third, when AI handles routine tasks like manual transcription, journalists lose the deep engagement with source material that traditionally built professional expertise. As one freelance journalist noted, ”Going over an interview with a fine-tooth comb and typing out every word really gets you inside it in a way that reading over a transcription and just checking for errors doesn’t.”
Beyond individual workflow challenges, the research uncovered deeper organizational tensions. Technical teams focus on usage statistics and efficiency metrics, while journalists prioritize editorial autonomy and journalism quality. These different success criteria create implementation conflicts that appear technical but actually reflect competing values about journalism’s purpose. One particularly interesting finding revealed that a developer had worked as a consultant for a media company for several years without ever speaking to a journalist in the organization.
For Finnish media specifically, language barriers compound these challenges. Participants consistently reported that AI tools perform excellently in English but poorly in Finnish, forcing workflow decisions based on technical limitations rather than editorial needs. When one journalist tested an AI system by asking it to find Finnish words ending with a specific syllable, the system provided ten responses – none of which were actual Finnish words.
This technological divide means journalists working in smaller language markets face persistent disadvantages, raising questions about equity in global AI development. Finnish media organizations have responded by developing internal AI solutions and blocking their content from training public AI models, creating a divergence where internal tools work well but popular systems like ChatGPT remain problematic for Finnish content.
The research identified four principles for ethical AI integration that build upon existing journalistic values rather than requiring entirely new frameworks: maintaining human oversight of all AI-generated content, ensuring environmental responsibility in AI use, preserving journalism quality over efficiency gains, and maintaining transparency with audiences about AI use.
Moving forward, the study recommends that media organizations establish permanent cross-functional teams bringing together editorial and technical staff, map existing workflows before selecting AI tools, and develop structured training programs for AI verification skills. Successful AI implementation requires addressing fundamental differences in how technical and editorial teams define success.
As Finnish newsrooms continue navigating AI adoption, this research suggests the most important question isn’t whether these tools can make journalism faster, but whether they can make it better while preserving the democratic accountability functions that define the profession. The efficiency paradox serves as a crucial reminder that in the rush to embrace AI, the hidden costs of implementation deserve as much attention as the promised benefits.
Media-alan tutkimussäätiö on rahoittanut Tomi Fischerin hanketta ”Managing AI in Media: Enhancing Journalist Workflow at Viestimedia”.
Maisterintutkielma Aalto yliopiston Aaltodoc-julkaisuarkistossa.