by Kyra Usielski
Artificial intelligence (AI) has already firmly established itself in our everyday lives and also in editorial offices and the media. While these technological innovations have the potential to positively influence journalistic work and possibly also the quality of publications, they also harbor risks, especially with regard to journalistic ethics. Recent academic publications shed light on how and for what purposes AI is currently being used in journalism, what opportunities and risks are associated with it and how this could develop in the future.
As a branch of computer science that “imitates human cognitive abilities” in order to process tasks independently, artificial intelligence has the potential to simplify editorial processes, but this raises considerations regarding both social and (media) ethical standards. This requires a critical interdisciplinary examination and systematization of the field of research – a “new media ethics” that deals with the ethical and social challenges of artificial intelligence.
Efficient paths from research to distribution
Learning algorithms are already supporting journalists in various phases of editorial processes by making them faster and more effective and thus creating more space for research and production of topics that cannot be automated in their complexity. (Elmer, 2022, p. 346 – 351)
The areas of application of algorithm-controlled systems in journalism can be divided into assistive, generative and distributive technologies. Assistive technologies support information retrieval, while generative technologies can generate texts. Distributing technologies, on the other hand, enable the personalized distribution of content. (Graßl et al., 2022, p. 3-10)
They facilitate access to information during research through continuous scanning and pattern recognition in online sources. Algorithms also enable the automated indexing and digitization of content, which in turn simplifies the analysis of complex source collections. They also play a crucial role in the verification of statements and sources by identifying and classifying digital statements. (Graßl et al., 2022, pp. 11-17)
AI systems play a central role in production by making everyday journalistic work more efficient through automated transcription of audio files, translation services and the generation of synthetic media products such as short texts on stock market, sports, weather and traffic news. Learning algorithms are also finding their first applications in moderation, even if their integration is still the subject of intensive research. In distribution, on the other hand, algorithms control the control of paywalls, enable a personalized customer approach and identify subscribers with a risk of cancellation. (Heesen & AG IT-Sicherheit, Privacy, Recht und Ethik, 2022, p. 11)
Between potential and challenges
Through qualitative interviews with experts from the fields of software production, editorial organization and media ethics, Graßl, Schützender and Meier (2022) attempt to create a knowledge base about the challenges and potentials of integrating artificial intelligence into editorial processes. The survey results show that AI in journalism offers potential, particularly in the areas mentioned in the previous section, but that successful integration is hampered by challenges in financing, training and corporate culture.
In addition, editorial effects can be identified, which are made up of organizational challenges, a lack of AI strategies and questions about the technical infrastructure. Journalists are therefore faced with the need to realign their skills, with technical skills in particular gaining in importance. There is also a risk of overwork as simple routine tasks are replaced by permanent, mentally demanding work.
According to Stefan Grill (Head of Innovation, 3pc GmbH) and Cécile Schneider (Product Lead BR AI + Automation Lab), however, the main risks lie in data protection, data quality, integrity and credibility. The media ethics debate therefore emphasizes both interdisciplinary collaboration and journalistic responsibility as a gatekeeper. In addition, the role of learning algorithms in journalism is clearly defined as assistance, in which the sovereignty of the relevance decision remains with the journalists.
The future of AI-driven journalism
Founded in 2020, Bayerischer Rundfunk’s AI + Automation Lab aims to critically evaluate AI in journalism through interdisciplinary collaboration and to develop editorially and ethically acceptable solutions in order to preserve the diversity of content and avoid undesirable side effects such as filter bubbles and harmful consumer behavior.
Overall, clear ethical guidelines, a critical approach, editorial monitoring, transparent disclosure and strategy in dealing with AI are crucial for the conscientious use of learning algorithms in order to be able to guarantee media trust and the democratic role of journalism in the future as well.
This is why the European Union has now also passed the first law on the regulation of artificial intelligence. The AI Act was adopted by the 27 EU member states on May 21, 2024 and is the world’s first comprehensive set of AI regulations. It aims to strengthen trust and acceptance of AI technologies by establishing clear rules for their use AI systems should be transparent, comprehensible, non-discriminatory and monitored by humans. EU member states are obliged to transpose the AI Act into national law. In Germany, Federal Digital Minister Volker Wissing emphasizes that maximum freedom for innovation should be guaranteed during implementation in order to promote competitiveness. The German government is also supporting the research and application of AI through various measures, including the establishment of AI service centers for science and industry.
This inevitably requires a “new media ethic” and a reorganization of education and training in order to enable the responsible use of AI systems and ensure the provision and transfer of technical expertise. Under these conditions, the AI-human division of labor not only promises to increase efficiency but can also contribute to improving the performance of the media landscape.
This article, Digital editors: Al in the media landscape, was first published by The European Journalism Observatory on July 8th