In January the EUSJA webinar “How to Utilise AI Tools in Daily Work Efficiently” welcomed Isabel Lerch, a data journalist from Germany’s public broadcaster NDR. She guided participants through the AI landscape that is reshaping modern journalism.
Why AI matters for journalists
Lerch began by reminding the audience that AI is no longer a novelty. It is now a daily assistant for research, fact-checking, writing, and even visual design. She stressed that AI can speed up routine tasks, allowing reporters to focus on deeper investigation and storytelling. But she warned that AI can also produce plausible-looking but false information (known as ‘hallucinations’), so human oversight remains essential.
The “big four”
According to Lerch, most journalists gravitate toward a handful of large AI services.
These include:
- ChatGPT (OpenAI): Easy to use, versatile, strong at generating drafts and brainstorming ideas.
- Google Gemini: Excellent at factual retrieval, offers a “deep research” mode that spends extra time gathering reliable sources.
- Claude (Anthropic): Writes with a more thoughtful, empathetic tone; good at coding and technical explanations.
- Mistral: European focused, respects EU data privacy rules, increasingly competitive on quality.
She noted that while many smaller tools exist, they often overlap with these four and rarely add unique value.
Prompt-writing
A large part of the webinar focused on how to talk to AI effectively. Lerch presented a “prompt formula” that she calls a package: each prompt should contain several ingredients.
- Persona – Tell the AI who it should act as (e.g., “You are an experienced science journalist”).
- Assignment – State the exact task (“Write a balanced overview of AI tools for journalists”).
- Context – Provide background information the AI needs (topic details, target audience).
- Format – Specify the output style (list, table, short paragraph).
- Goal – Clarify the purpose (“Help me decide which tool to try first”).
- Extra Details – Any additional constraints (word limit, tone).
Using this structure reduces ambiguity and limits the aforementioned hallucinations. Lerch demonstrated the method with a real example: asking an AI to list experts on electric-vehicle policy and receive the results in a tidy table.
Practical use-cases for journalists
Lerch illustrated several everyday scenarios where AI can save time:
- Finding perspectives – By feeding a controversial topic (e.g., wolf-population quotas in Germany) into an AI, she received a list of stakeholder groups that she had not considered, broadening her coverage.
- Preparing interview questions – A generic prompt can generate a set of tailored questions, which she then refines for each interviewee.
- Locating interview subjects – Using AI to scan public sites such as Reddit, she identified personal stories (e.g., a single mother facing eviction) and extracted concise summaries and links.
- Summarising messy notes – After a phone interview, she copies her handwritten notes into an AI and receives a clean, readable summary.
- Creating visual aids – Some AI tools can turn raw data into charts or even draft PowerPoint slides, speeding up the production of multimedia stories.
Pitfalls and ethical considerations
While enthusiastic, Lerch reminded participants of the risks, such as data privacy, copyright, and bias and misinformation in the results. Journalists should also be wary of tool dependence; over-reliance on a single platform can create a lock-in effect, limiting flexibility.
She encouraged journalists to treat AI as a helper, not a replacement, and to keep a critical eye on every output.
Looking ahead
Lerch concluded that AI will continue to evolve, but the core mission of journalists to provide accurate, fair, and insightful storytelling will stay the same. She urged her peers to experiment, share experiences, and teach newcomers how to harness AI responsibly.