Using AI in your media: what questions should you be asking?

Gérald Holubowicz is a journalist, teacher and media consultant specialising in synthetic media. He is the creator of synthmedia.fr, a medium that pays close attention to what tech and AI are changing in our lives. Artificial intelligence is indeed an integral part of our lives through tools such as waze, google map… AI is not something really new… and the media world is no exception. For instance, Le Monde or other newsrooms use it to generate some content for elections or sports results.

Three figures to start with

In the report, Generative AI in journalism / The evolution of newsworks and ethics in a generative information ecosystem, there are three key figures.

• 80% of respondents say they are familiar with generative AI, the basic knowledge seems to be there, people seems to understand what is generative AI.

• 73% use generative AI, that could cover wide range of practices: using chat gpt to summarise a text or whatever. 

• 70% use it for the production of textual content. What is a very high number.

Different uses in the press

Many media are interested in using AI and are already using it to produce content. Like newsletters, or even summaries of comments on articles, when there are a lot of them. This is the choice made by the sports daily L’Equipe, for example. Le Monde, too, has created a partnership with open IA, to ensure the reliability of the responses from its ChatGPT tool.

But not all uses are successful…

Sept à huit magazine used AI to open their news mag story, to illustrate Antwerp harbour. People who know Antwerp are able to say if yes or not this the real port. Looking at the images, for example, you realise that the uniform of the cop is not accurate. Was such a use necessary? Another problematic case: images generated by AI ended up in the Adobe store despite the existence of an “IA generated image” credit.

Click here to watch the masterclass video:

Questions to ask

According to Gérald Holubowicz, we must not lose sight of several questions: Is AI a solution for everything? How can we integrate them? For example, thinking about a new organisation? Can we do differently? The main ethical concern remains the lack of human supervision. That’s probably the main question. But there are still a lot of issues around risk, such as how to manage bias? How to manage feelings of downgrading inside the newsrooms? How can we build trust?

And uses to be explored

According to Gérald Holubowicz, AI can be used primarily for broadcasting and creative assistance through data analysis, transcription or development. But it can also be used for helping you to think or for promotional support. This is particularly the case for personalising content or improving headlines. Our contributor goes into more detail in his book and in the video. 

Note: L’AI dans les Rédactions by Gérald Holubowicz, a guide for newsrooms and content creators wishing to implement tools based on artificial intelligence systems within their newsrooms.