MELLAN MÄNNISKA OCH MASKIN. Yrkesverksammas upplevelser av generativ AI i kommunikativt arbete

Sammanfattning

Generative artificial intelligence has quickly become part of everyday work in media andcommunication professions. Tools such as ChatGPT, Gemini, and Copilot are now used foridea generation, text drafting, research, and strategic support, tasks that have traditionallyrelied on human expertise and judgement. Unlike earlier digital tools, generative AI producescontent that can appear confident and well-structured, even though it lacks understanding orresponsibility. This creates new ethical challenges for professionals who remain accountablefor quality, accuracy, and credibility. At the same time, organisational policies and formalguidelines often lag behind technological development, leaving individuals to handle ethicalquestions on their own.Against this background, this study explores how professionals within media andcommunication talk about their use of generative AI and how they experience its role in theirdaily work. Rather than examining AI use from a technical perspective, the focus is onpractitioners own descriptions, reflections, and concerns. Special attention is given to howethical boundaries are perceived and how questions of responsibility, transparency, andprofessional integrity are handled when AI becomes part of creative and strategic workprocesses.The study addresses two research questions:1. What ethical limitations and boundaries do professionals experience when usinggenerative AI?2. How do they handle issues of responsibility, transparency, and professional integrityin relation to this use?The study is based on a qualitative research design using semi-structured interviews witheight professionals working in journalism, strategic communication, marketing, and publicrelations. All participants had experience using generative AI in their professional work,which was a deliberate choice to ensure informed and reflective accounts rather thanspeculative opinions. The interviews focused on how participants describe their AI use, howthey reason about ethical concerns, and how they understand responsibility and transparencyin AI supported work. The material was analysed using thematic analysis, allowing recurringpatterns and shared perspectives to be identified across the interviews.The findings show that generative AI is widely integrated into everyday work practices, butthat its use is rarely seen as straightforward or ethically neutral. Rather than relying on formalrules or clear organisational policies, participants describe how ethical boundaries arehandled through personal judgement and situational decision-making. Responsibility isconsistently understood as remaining with the human user. Even when AI plays a significantrole in shaping ideas or content, participants emphasise that they are fully accountable for outcomes and consequences. AI is mainly described as a supportive tool, not as anindependent decision-maker.The study also shows that transparency around AI use is not viewed as an absolute principle.Instead, decisions about disclosure are negotiated depending on context, audienceexpectations, and professional norms. Being transparent is often seen as important formaintaining trust, but participants also describe situations where AI use is treated as a routinetool and therefore not explicitly communicated.Overall, the study highlights that ethical governance of generative AI largely takes place ineveryday practice rather than through formal regulation. Professional experience, judgement,and norms play a key role in managing responsibility, transparency, and integrity. Byfocusing on practitioners own accounts, the study challenges assumptions that professionalslack awareness of AI-related risks. Instead, many users actively reflect on ethical boundariesand deliberately position AI as support rather than a replacement for professional judgement.At the same time, this places increased responsibility on individuals, which may create newpressures in increasingly technology-driven work environments.