Can artificial intelligence sense the stress of a leader who’s speaking in front of a room full of journalists, executives, employees or activists? Can it notice how he might fidget with his tie nervously, feeling it’s too tight? How she discreetly wipes her sweaty palms? Can artificial intelligence, at the end of a meeting, approach a source to ensure it correctly understood some key bits of information, has the proper context and background, and that there’s not another angle to pursue?

No. Programmed based on past data, artificial intelligence cannot beat a journalist on their own turf. Especially when journalists leave the newsroom and put down their screens to be where they should be: on the ground, engaging with sources and stakeholders.

However, AI, like all the electronic tools we’ve been using since the 1990s, can certainly ‘augment’ a journalist.

Take one of the examples we’re working on to improve our newsroom processes: Journalist A needs to go to Company B for an annual results presentation. Instead of waiting for the official announcement that might or might not have relevant comparisons with previous years, previously we had to manually enter each number into an Excel spreadsheet. With ChatGPT or similar tools, we could quickly pull together our own prep sheet. We currently are identifying several other use cases to help us work more efficiently and effectively.

We have been testing and will continue to test different solutions that provide practical support and align with professional standards. Our use of AI will follow several principles, first and foremost being that no content will be published without being authored and edited by a real-life professional journalist.

Every published article includes a byline for one or more journalists, who remains accountable for the truth and relevance of the information in the story. AI tools are tools, but in no way information sources.

Images

Delano and Paperjam use professional photographers, including those on staff and a pool of freelancers. We have paid subscriptions to illustrative picture databases, who in turn pay photographers for their work. In addition, companies and organisations provide us with portraits and images based on interviews or specific requests. We will continue to do so.

Only as a last resort would we use an image entirely generated by artificial intelligence--with a clear explanation of how the image was produced--unless it is to illustrate a phenomenon related to artificial intelligence. These images must directly relate to the content and be properly labelled.

Ethical standards

There is nothing about artificial intelligence that changes the codes of ethics that journalists follow. These include the , , Luxembourg’s on freedom of expression in the media and the Luxembourg Press Council’s , updated in November 2022.

AI guidelines

Delano and Paperjam will:

- Publish original articles written by human journalists.

- Ensure that human journalists are responsible for the accuracy, fairness, impartiality and quality of their stories.

- Use AI as a research and support tool, such as: data, text and image analysis; transcription, translation and proofreading; and headline and social media post suggestions.

- Vet and verify AI-generated content, reviewing for potential bias, inaccuracies, misinformation, intellectual property breaches and compliance with ethical standards.

- Be transparent about our use of AI tools.

Delano and Paperjam will not:

- Use AI to replace human journalists.

- Directly publish AI-generated text or photorealistic images (unless AI is the focus on the story, this is clearly disclosed and errors are flagged).

- Use AI tools or AI content as a primary source.

- Input confidential or sensitive data into AI engines that cannot safeguard the information.

Read our AI policy in French on the site