AI in Newsrooms

What ethical questions should be asked to ensure responsible implementation?

After a brief overview of the place of AI in the press, this guide addresses a series of questions designed to help information professionals think about the use of AI.

The main aim is to provide a number of ethical insights to help journalists and editors who wish to draw up a charter to govern the use of generative AI or, more broadly, artificial intelligence systems within newsrooms.

More broadly, this guide provides a starting point for an industry-wide conversation on digital ethics in the press.

What you'll get in this guide?

After an OVERVIEW that explores the way in which existing charters are drafted, some of the points put forward by these texts are explored, in particular, the question of supervision, transparency, data protection, environmental issues, and the question of their durability and "hallucinations". The first part concludes with a set of reasons in favor of an ethical and responsible approach to AI by publishers.

The second part, entitled IMAGINE A MODEL, looks at the scenarios in which solutions based on artificial intelligence systems could be developed or implemented in newsrooms.

QUESTIONS TO ASK brings together a set of questions for each type of audience (audience, editorial staff, support functions) to create a charter for the use of ethical and responsible artificial intelligence systems.

IS PRESSGPT POSSIBLE? explores the subject of creating finely-tuned language models for the newspaper industry using media archives and based on an open-source model overseen by a not-for-profit organization.

Finally, you will find a CHART EXAMPLES and a GLOSSARY of common terms to help you better understand the subject and complete your knowledge.

Updates

Version 1.0 (04/09) : English version published