They Think AI Can Do More Than It Actually Can: Practices, Challenges, & Opportunities of AI-Supported Reporting In Local Journalism
Through interviews with 21 German local journalists, this study reveals that while newsrooms are increasingly adopting AI to address revenue declines, journalists currently underutilize its potential due to limited awareness of its capabilities, highlighting specific challenges and opportunities for improving AI-supported reporting through a socio-technical lens.
Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
Imagine a local newsroom as a small, busy kitchen in a neighborhood. The chefs (journalists) are responsible for feeding the community fresh, accurate information every day. But lately, the kitchen is running low on ingredients (money), the staff is shrinking, and the menu is getting harder to prepare.
Enter AI: a new, high-tech sous-chef that promises to chop vegetables, mix sauces, and even write recipes at lightning speed.
This paper is a report from two researchers who went into these local kitchens in Germany to ask the chefs: "How are you actually using this new robot helper? Is it a miracle worker, or just a fancy gadget gathering dust?"
Here is the story of what they found, explained simply.
1. The Big Misunderstanding: "The Magic Wand" vs. The "Scalpel"
The researchers found that many local journalists have a bit of a mismatched expectation.
- The Myth: They think AI is a magic wand that can instantly solve all their problems, find hidden secrets, and write perfect stories.
- The Reality: Currently, AI is more like a very fast, but sometimes clumsy, assistant. It's great at chopping onions (summarizing text, fixing grammar, brainstorming headlines), but it's not great at cooking the main course (analyzing complex data, finding deep truths, or understanding local nuance).
The paper's title, "They Think AI Can Do More Than It Actually Can," is like a parent realizing their child thinks a toy robot can actually fix the car. The robot can beep and flash lights, but it can't change the oil.
2. The Chefs' Current Habits: "The Text-Only Zone"
When the researchers asked how the journalists use AI, the answer was mostly about words, not numbers.
- What they do: They use AI to fix typos, write catchy headlines, translate articles, or summarize long documents. It's like using a spell-checker on steroids.
- What they don't do: They almost never use it to dig through piles of raw data (like police reports, election stats, or hospital records). Why? Because they feel like they are trying to drive a race car without a license. They lack the "data literacy" (the skills to read and understand numbers) and feel intimidated by spreadsheets.
The Analogy: Imagine a chef who is amazing at plating food beautifully but is terrified of the stove. They will use AI to arrange the garnish, but they won't let the AI cook the steak because they don't trust it not to burn the house down.
3. The Hurdles: Why They Hesitate
The journalists listed several reasons why they aren't letting the robot take over the data work:
- The "Black Box" Fear: They don't know how the AI gets its answers. If the AI makes a mistake, the journalist is the one who gets fired, not the robot.
- The "Hallucination" Problem: AI sometimes makes things up with total confidence. In journalism, making things up is a cardinal sin.
- The "Local Context" Gap: AI knows everything about the world, but it knows nothing about your town. It doesn't know that the "scary" neighborhood is actually just a construction site, or that a specific politician has a history of lying about a specific local issue.
- The Time Trap: Ironically, checking the AI's work sometimes takes more time than doing it yourself, because you have to fact-check every single thing it says.
4. The Golden Opportunity: "The Super-Powered Reporter"
Despite the fears, the journalists are excited about what AI could do if it were designed better. They don't want AI to replace them; they want it to be a super-tool.
Here is their dream scenario:
- The "Data Detective" Assistant: Imagine an AI that can read 1,000 police reports in 10 seconds, find the pattern that "burglaries are up 20% in the north district," and draw a map for the journalist.
- The "Time-Saver": If the AI handles the boring stuff (sorting data, transcribing interviews, checking basic facts), the journalist can spend their time doing what humans do best: talking to people, investigating deep stories, and understanding the human heart of the issue.
The Metaphor: They want the AI to be the mining drill that digs up the gold (data), so the journalist can be the jeweler who turns that gold into a beautiful necklace (a story).
5. The Recipe for Success: What Designers Need to Do
The researchers conclude that if we want AI to help local news, we need to stop building "one-size-fits-all" tools and start building tools that respect the local chef's kitchen.
- Don't just give them a hammer; give them a guided drill. The tools need to explain why they found a pattern, not just show the result.
- Teach them to fish. Journalists need training to feel confident with numbers, not just words.
- Keep the human in the loop. The AI should be a co-pilot, not the captain. The human must always hold the controls, especially when it comes to ethics and local truth.
The Bottom Line
Local journalism is the "watchdog" of the community, keeping an eye on power and informing neighbors. If the watchdog gets tired and stops barking, the neighborhood gets unsafe.
This paper argues that AI can help the watchdog stay awake, but only if we stop expecting the robot to be a human and start building tools that help humans do their best work. We need to bridge the gap between the "tech wizards" and the "storytellers" so that local communities don't end up in the dark.
In short: AI is a powerful new engine, but local journalists are still learning how to steer the car. We need to build better maps and training wheels, not just faster engines.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.