Don’t Blindly Blame the Bots: How AI Is Helping Journalism

Let’s face it: artificial intelligence (AI) has become the punching bag of choice in parts of the Brussels bubble, particularly in press-related discussions. From public hearings to position papers, AI is regularly paraded around as the looming villain behind layoffs, eroded editorial integrity, and the presumed death of quality journalism. But step away from the microphone theatrics and peek inside actual newsrooms, and a very different story unfolds, a story far richer, more nuanced, and, dare I say, far more hopeful.

AI as the Newsroom’s Secret Weapon: When AI Helps Journalism Get Its Groove Back

The everyday reality in many pressrooms, from sprawling international giants to scrappy regional outfits, is that AI is less a grim reaper and more an underpaid assistant who never sleeps. Automated transcription tools, machine-learning-powered investigative aids, translation software, and content repurposing platforms are not just theoretical perks: they’re already reshaping how journalists work, and crucially, where they can add value.

Table 1: Core Journalistic Tasks Augmented by AI and Their Positive Outcomes

AI Application Area	Specific AI Capabilities/Tools Mentioned	Positive Outcome for Journalism
Research & Discovery	LLMs for sifting data, AI for analysing large datasets (e.g., interviews, public records)	Time savings for in-depth reporting, identification of new story angles, ability to tackle complex investigations
Content Drafting & Editing	AI editing assistants, generative AI for initial drafts, headline/SEO optimisation tools (e.g., Producer-P)	Faster content production, improved clarity and conciseness, enhanced reach through optimisation
Data Analysis	AI for identifying patterns, trends, and anomalies in complex data (e.g., financial, health, environmental)	Deeper insights, evidence-based reporting, and uncovering stories previously hidden in data
Transcription & Translation	Automated transcription services, AI-powered translation tools	Significant time savings, increased accuracy over manual methods, and broader audience reach through multilingual content
Fact-Checking Assistance	Experimental AI systems for automating routine verification tasks	Potential for improved efficiency and accuracy in verification, freeing journalists for more complex fact-checking
Content Transformation & Format	AI tools to convert text to audio/video, create summaries, and personalise content delivery	Wider audience engagement through multiple formats, content more accessible and relevant to diverse audience preferences

Mediahuis in Europe is strategically embedding AI tools into its new editorial system, CUE. This integration aims to support daily workflows, improve operational processes, transform content into new formats like audio and video, and ultimately “keep our model afloat” in a challenging media landscape.

The New York Times (U.S.) has leveraged AI for complex investigative journalism. AI’s capacity to analyse extensive interview footage and hundreds of Zoom calls proved indispensable for developing in-depth exposés—tasks previously deemed “physically impossible” due to sheer data volume. Furthermore, their “keyword disappearance project” utilised machine learning and AI to track alterations on government websites, moving beyond simple keyword searches to identify ideologically motivated changes, even across different languages.

The Financial Times (U.K.) has fostered innovation through its “AI playground,” an internal tool allowing staff to safely experiment with Large Language Models (LLMs) using existing content. This initiative encourages learning and discovery. On the audience-facing front, the FT employs AI to generate discussion prompts within articles, successfully boosting comment section interaction and, consequently, subscriber retention. They also provide AI-generated bullet-point summaries at the top of articles, a feature developed in response to reader behavior (pasting content into ChatGPT for summaries). This initiative has maintained high factual accuracy and has not negatively impacted audience engagement, with Liz Lohn noting, “We have even seen a little bit of a positive impact on overall engagement”.

The case of iTromsø, a small Norwegian newsroom, and its AI tool “Djinn” is particularly insightful. “Djinn” is an AI-powered data journalism interface designed to enhance newsgathering, analysis, and story summarisation. Crucially, iTromsø journalists participated in training the AI, leading to high user adoption because the system was “well-explainable and built around users”. Journalist and data scientist Nikita Roy lauded Djinn as “one of the best use cases of AI deployment in the service of journalism,” highlighting its ability to help smaller newsrooms uncover original stories despite scarce resources. This demonstrates that AI’s benefits are not exclusive to large, well-funded organisations; smaller players can also innovate effectively.

Other notable examples include:

• Hearst Newspapers (U.S.) with “Producer-P,” a Slack-based tool using GPT models to assist journalists in creating optimised headlines, SEO titles, URLs, related links, and notification summaries across multiple newsrooms.
• Der Spiegel (Germany) is experimenting with an AI fact-checking system to automate routine verification, aiming to improve efficiency and accuracy while upholding journalistic integrity.
• Zamaneh Media (Netherlands), a small Persian-language newsroom, developed AI tools—Newsletter Hero and Samurai—that dramatically streamlined newsletter creation and the translation of long articles, significantly boosting efficiency for a two-person team.
• The BBC World Service (U.K.) successfully used AI to sift through vast amounts of open-source data for an investigation into a Russian military unit, an effort that won the Online Journalism Award for Excellence in AI Innovation.
• The Times of India (India) built “Signals,” an in-house AI recommendation system that personalises content for its vast readership based on user preferences, news trends, and engagement data, thereby transforming readers’ interactions with news content.
• Tamedia (Switzerland) and Russmedia (Austria) have implemented comprehensive AI training programs and strategic integration initiatives, such as Tamedia’s “AI Pyramid” and Russmedia’s “AI Future Team.” These efforts have led to reduced skepticism, an increased number of AI enthusiasts within their companies, and daily AI use by many employees, underscoring the critical role of change management and education in successful AI adoption.

Table 2: Global Newsroom AI Initiatives and Their Positive Contributions

News Organisation (Country)	AI Initiative/Tool & Brief Description	Primary Benefit Category
Mediahuis (Europe)	Integration of AI tools into CUE editorial system for process improvement, content transformation (text to audio/video).	Efficiency, Content Diversification, Sustainability
The New York Times (U.S.)	AI for analysing extensive interview footage, Zoom calls for investigations; ML/AI for tracking website changes beyond keywords.	Investigative Power, Data Analysis
Financial Times (U.K.)	“AI playground” for internal experimentation; AI-generated discussion prompts; AI-generated article summaries.	Audience Engagement, Efficiency, Staff Upskilling
iTromsø (Norway)	“Djinn” AI-powered data journalism interface for newsgathering, analysis, and summarisation; journalists involved in AI training.	Data Journalism, Efficiency, Empowering Small Newsrooms
Hearst Newspapers (U.S.)	“Producer-P” Slack-based tool (GPT-4, GPT-3.5-Turbo) for headlines, SEO, URLs, related links, and notification summaries.	Efficiency, Content Optimisation
Der Spiegel (Germany)	Experimental AI fact-checking system.	Fact-Checking, Efficiency
Zamaneh Media (Netherlands)	“Newsletter Hero” and “Samurai” AI tools for newsletter creation and translation (Persian to English).	Efficiency, Translation, and Empowering Small Newsrooms
BBC World Service (U.K.)	AI tools for sifting vast amounts of open-source data for investigations (e.g., Ukraine war reporting).	Investigative Power, Data Analysis
The Times of India (India)	“Signals” in-house AI recommendation system for real-time news personalisation.	Content Personalisation, Audience Engagement
Tamedia (Switzerland)	“AI Pyramid” strategy, AI Lab, training programs, AI hackathons; integrated employees in development.	Staff Upskilling, Innovation Culture, Change Management
Russmedia (Austria)	“AI Future Team,” surveys, custom tool development, 70+ workshops, and prompt catalogs.	Staff Upskilling, Efficiency, Change Management

These examples collectively paint a picture of an industry actively and successfully harnessing AI to enhance its capabilities and better serve its audiences.

So why the disconnect between what’s said in the Berlaymont and in European Parliament Committees vs. what’s happening on the editorial floor?

The Brussels Disconnect: Regulation vs. Reality

The EU, with all its procedural splendor, does love a good cautionary tale. But when it comes to AI in journalism, we seem to be watching a rerun of the same old dystopian drama, despite a growing volume of evidence suggesting the sky is not falling.

Yes, AI can create problems. And yes, the industry needs robust ethical frameworks, transparency, and critical human oversight. But that’s a separate discussion from the one where AI is painted as the main culprit for layoffs, pay freezes, and the devaluation of journalistic labor. The real villain here—spoiler alert—is decades of cost-cutting, shrinking ad revenues, and a stubborn reluctance to reinvent outdated business models. AI just happens to be the convenient fall guy.

Rather than hollowing out journalism, AI is often revitalising it—especially in corners of the world where resources are thin and deadlines tighter than a eurocrat’s calendar. This is not about robots doing journalism. It’s about journalists doing better journalism—faster, deeper, and with fewer headaches.

A Call for Constructive Conversations (and Less Scaremongering)

If the Brussels crowd truly wants to support a robust, independent press, the conversation needs a serious recalibration. Let’s move away from the oversimplified AI-as-threat narrative and toward a recognition that AI, when handled responsibly, is an asset, not an adversary.

We also owe it to journalism (and to public discourse more broadly) to stop blaming a set of algorithms for decisions made in boardrooms. If staff are being cut, let’s have the courage to say it’s about margin management, not machine learning.

AI is not the problem. It’s a tool. The problem lies in how we choose to use it or misuse the story around it. Let’s not confuse political grandstanding with journalistic reality.

In the press sector, AI’s integration is not just a technical upgrade; it’s a strategic one. And when done right, it doesn’t replace the journalist. It just makes them better at what they already do best: uncovering truths, connecting the dots, and keeping the powerful accountable.

The press doesn’t need a savior, but it also doesn’t need another fall guy. What it needs is space to adapt, tools to thrive, and policy conversations that reflect reality, not just regulatory paranoia or lobbynomics.

Written by Caroline De Cock, LL.M. , Head of Research.