09 October, 2020
Experts emphasise the great benefits of using AI in journalism but highlight the need for heavy investments and an ethics-by-design approach
The last focus session of the German EU Presidency related to Media in the Digital Society, titled "How can AI contribute to ensuring quality journalism?" took place on 1 October 2020. The session was organised in cooperation with the Deutsche Welle, the German international broadcaster and invited four AI experts, Kenza Ait Si Abbou Lyadini (Deutsche Telekom IT), Meredith Broussard (Arthur L. Carter Journalism Institute of New York University), Charlie Beckett (London School of Economics) and Rasmus Nielsen (Reuters institute) to discuss the advantages and risks of using AI in journalism.
AI, an ally of the journalist?
According to the experts, there is a need to ‘demystify’ Artificial Intelligence as it does not just belong to the robotic world but is also a tool used by all in the daily life. AI can be a real ally for the journalist if used for the right task, such as monotonous and repetitive tasks, fact-checking, sorting out the huge quantity of information etc. In that respect, AI may enable journalists to focus on more qualitative tasks while creating new kinds of jobs as more skills and profiles will be needed.
AI, a threat to journalism?
Products stemming from AI tools inevitably embed some unconscious bias of the small group of people who created them. Such small groups of people do not necessarily reflect on the reality of society - especially in terms of diversity of people and ideas - and this can be a threat to democracy - especially when AI is dominated by a few major companies. The experts hence called for more regulation and for an ethics-by-design approach. The impact of AI and ethical issues should be addressed by everyone involved: from the tech world to the editorial managers, the journalists and the regulatory authorities.
AI and quality journalism: where are the limits?
The technical obstacle: whereas AI tools can help journalists (through fact-checking, correction of editorial bias…), it cannot replace the content analysis that only a journalist can provide.
The capacity to invest: implementing AI tools in newsrooms requires to rethink the strategy and editorial work in general (not just a metter for the IT department) and to support heavy investment without guarantee of results. As Meredith Broussard put it, between ‘good’, ‘fast’ and ‘cheap’, you can only pick two.
Source: Germany’s Presidency of the Council of the EU
Further EPRA Background: AI and machine learning were the focus of a plenary session at the 50th EPRA meeting in Athens. On that occasion, EPRA Vice-Chair Oliver Gerber produced an in depth background paper. The impact of algorithms is addressed this year in the annual work programme from the perspective of 'media plurality' (See EPRA podcast Season 1) and 'transparency and trust' (See upcoming EPRA podcast - Season 2).