Can you tell the difference between AI-generated and real images? Do you feel left behind when it comes to optimising your workflow with the latest AI tools? During the two-day seminar, participants delved into these and other relevant questions concerning the role of journalists in an era of disinformation and enhanced digital tools.
The workshop covered two key areas. The first part focused on identifying AI-generated content, while the second provided participants with useful tools and tips for the use of AI, such as chatbots, in their daily work.
During the two meetings, the European Newsroom greeted around 40 journalists from ten participating agencies. They were dialling in from Sweden and Spain, from France and Belgium and several other countries across Europe.
Participants from the workshop smiling after the closing remarks from Christian Röwekamp. Photo: enr
PART I: Unmasking AI-Generated Content
On the first day of the workshop, Stefan Voss and Patrick Neumann prepared several interactive exercises involving examples of AI-generated videos and images. These included remakes of real celebrities such as Dua Lipa, as well as key figures such as Pope Francis and US President Donald Trump – but also real videos like a volcano eruption.
Participants were then taken to breakout rooms to discuss the veracity of the images applying these techniques. Finally, some AI-detection tools were shared, such as thehive.ai or Image Whisperer, and AI-video tools like Sora.
Agenda and one AI-created photo activity featuring the former Pope Francis on November 6. Photos: enr
PART II: How to use AI in your job
After learning how to detect AI-generated content, the next session focused on implementing AI-enhanced tools at work. During the first two hours, participants learnt how to reduce errors and increase precision using diligent, structured prompts, as well as how to foster research using powerful prompts.
Later, participants were invited to try out the prompts they had been given on their chosen AI platform – i.e. Gemini, ChatGPT, Claude or Perplexity – with the aim of reducing errors. Then, the facilitators explained how participants could increase their productivity by creating a custom assistant, such as a Custom-GPT, and by implementing data analysis tools and semantic search in their day-to-day workflow.
In the end, the participants agreed: the field of Artificial Intelligence is developing so rapidly that it is essential to keep up with it and engage in regular dialogue. In addition, it is beneficial that the enr provides a platform for organising such cross-border learning and exchange among news agency professionals.
Agenda and structured prompts explained by Stefan Voss and Patrick Neumann on November 10. Photos: enr