TV 2

AI Training Intake Analysis

A newsroom full of potential, curiosity and journalistic sharpness

Work smarter, not harder: AI for journalists
9-10 April 2026 20 participants Source: Google Forms

Twenty journalists from TV 2 Odense have openly and honestly shared their AI knowledge and learning ambitions. What stands out immediately: this is a remarkably engaged group. From Bo's pioneering deepfake work since 2018 to Jakob's vision of AI pattern recognition in legal data, every single participant brings something unique to the table. The range of experience is not a weakness but a strength — it means the advanced users can mentor, the beginners can ask the sharp questions nobody else thinks of, and everyone learns from the exchange. This group is ready.

Key Figures

6.0
Strong foundation
6
Solid starting position
3 – 9
Beginners to power users
1.7
Diverse = rich exchange

Individual Profiles

Bo Bergstedt

The deepfake pioneer and natural mentor
9 / 10
"Have been involved since I started doing deepfakes back in 2018. So don't customize the course for me — I am an outlier at TV2."

Personal Advice

Bo is a rare asset in any newsroom. While most journalists are still learning to use chatbots, Bo has been working with AI since 2018 — when deepfakes were barely on anyone's radar. His self-awareness is equally impressive: he calls himself an "outlier" and explicitly asks the training not to be calibrated to his level. That kind of generosity is what makes a team stronger. His goals reveal his true motivation: he wants to build tools that help his colleagues battle AI-generated misinformation, and his measure of success is not personal achievement but "seeing journalists getting more comfortable with AI." That is leadership. Bo can serve as a bridge between the technical possibilities and the editorial reality, helping translate complex AI concepts into practical newsroom workflows. His deep technical knowledge combined with his collaborative spirit makes him the ideal catalyst for the group.

Christian Jessen

The skeptical thinker with deep understanding
9 / 10
"I know that most AI-software are language models, so AI is usually great at generating and improving text."

Personal Advice

Christian demonstrates something exceedingly rare: he understands the fundamental architecture of AI systems. His knowledge that these tools are "language models" — and his ability to derive practical conclusions from that understanding — places him head and shoulders above the average user. His healthy skepticism is not resistance; it is precisely the kind of critical thinking every newsroom needs. He uses AI for summarization but refuses to trust it for knowledge generation, because he has seen it fail. That is empirical reasoning, not fear. His interest in OSINT setups and his question about building a proper social media monitoring deck shows strategic thinking about information architecture. Christian could become the group's go-to person for verification workflows and outside-the-box thinking about how AI fits into investigative processes.

Pelle Lykkebo Mørk

The verification specialist who proved it works
8 / 10
"I've used AI extensively across research, drafting, live news production — and also worked on internal implementation, building prompts and workflows adapted to newsroom needs."

Personal Advice

Pelle is not just using AI — he is implementing it at the organizational level. He builds prompts and workflows adapted to newsroom needs, which means he understands both the technology and the editorial context. His verification of US E-3 Sentry damage using flight tracking data is a textbook example of how AI-assisted research can produce publishable, factual journalism. That he links to a real TV 2 breaking news article to prove it shows confidence backed by results. His concerns about editorial independence and data security are not obstacles — they are the guardrails every responsible newsroom should have. Pelle understands that AI works best for research overviews, framing, first drafts, and structured prompting. That clarity about what AI does well and where it falls short is exactly the kind of nuanced understanding that makes him invaluable as a workflow designer for his colleagues.

Emil Gjerding Nielson

The implementation champion building for others
8 / 10
"Working on implementing editorial AI tool for my fellow colleagues. Use it in my writing amongst other things."

Personal Advice

Emil is doing something truly commendable: he is not just learning AI for himself, he is building editorial AI tools for his colleagues. That dual focus — personal mastery and organizational uplift — is the hallmark of someone who thinks beyond their own desk. His goals are precisely targeted: automate feedback loops and improve writing quality. His definition of success — gaining "insight into tools, prompts etc. that I can use in my daily work" along with "reflections on possibilities" — shows someone who wants both practical skills and strategic vision. Emil is the kind of participant who will take what he learns in this training and multiply its impact across the entire newsroom. His implementation work means every technique he picks up has the potential to benefit dozens of colleagues.

Peter Møller

The versatile researcher with journalist's instinct
7 / 10
"I use AI for translation, summarization, verification, research — but the challenge is that AI invents sources."

Personal Advice

Peter has identified the single most important limitation of AI in journalism: hallucinated sources. The fact that he caught this, named it, and treats it as a challenge rather than a reason to stop using AI entirely shows remarkable maturity. He already uses AI across four critical workflows — translation, summarization, verification, and research — which places him firmly in the power-user category. His wish list is focused and practical: research in journalistic sources, scientific databases, and social media, plus image and video verification. Every one of these is a skill that will pay immediate dividends in his daily work. Peter's combination of broad experience and sharp critical awareness makes him an ideal participant who will get maximum value from the training.

Lars Apel

The agent builder scanning the world for stories
7 / 10
"Have tried setting up agents to help me scan news across the world."

Personal Advice

Lars is already operating at a level most journalists have not yet imagined: building AI agents. While others are still asking chatbots questions one at a time, Lars is constructing automated systems that scan global news for him. That is not just using AI — that is engineering AI workflows. His report that it "works for finding sources" means he has already validated the concept in practice. He works across CoPilot, ChatGPT, and is considering Claude — showing a healthy multi-tool approach rather than loyalty to a single platform. His ambition to have AI scan news and pitch stories represents the frontier of AI-assisted journalism. His success criterion — "getting a clearer picture of what is possible right now" — is the question of someone who has already started building and now wants to calibrate his ambitions against the state of the art.

David Buch

The daily practitioner hungry for depth
7 / 10
"I have taken a course on TV 2's summer workshop and use it almost everyday in my work as a reporter at tv2.dk."

Personal Advice

David embodies the ideal learning trajectory: he took a course, integrated the skills into daily practice, and now wants to go deeper. Using AI "almost everyday" as a reporter means it is already part of his professional toolkit, not an occasional experiment. His three learning goals — more confidence, better prompting, stronger critical sense — are precisely the areas where experienced users make the leap from competent to excellent. His explicit interest in "the pros AND cons of AI" signals intellectual honesty: he is not looking for cheerleading, he wants the complete picture. That balanced perspective will serve him well. David's daily practice means every new technique he learns in this training will be applied immediately and repeatedly, creating a powerful feedback loop of continuous improvement.

Sebastian

The pragmatic toolbox builder
6 / 10
"ChatGPT and other programs like it. Also some picture verification and generating tools."

Personal Advice

Sebastian has already moved beyond text-only AI into the visual domain — picture verification and image generation tools. That breadth is notable: many journalists at his level have only explored chatbots. His experience with visual AI tools means he already understands that AI is not just about text, which gives him a significant head start for verification work. His success criterion cuts straight to the point: "Just that I get some tools that I can use in my daily work." That is not a modest goal — it is a supremely practical one. Sebastian is the kind of participant who will immediately put new skills into action. With his existing foundation in both text and visual AI, he is perfectly positioned to become a well-rounded AI practitioner who can tackle any type of content that crosses his desk.

Anne Fuglsang Borg

The ethical thinker with advanced prompting instinct
6 / 10
"AI has worked best when I've been the most concrete regarding prompting."

Personal Advice

Anne has already discovered the single most important principle of effective AI use: specificity in prompting. Her observation that "AI has worked best when I've been the most concrete" is a breakthrough insight that many users never reach. She has applied AI to academic writing, brainstorming, data sorting, and grammar — a diverse and intellectually demanding set of tasks. What makes Anne truly stand out is her ethical compass: she wants to "implement AI without compromising ethical standards." In an era of uncritical AI adoption, that question deserves to be celebrated. Her desire to "gather more knowledge about AI in general" reveals intellectual curiosity that goes beyond mere tool-learning. Anne brings the rare combination of practical experience, principled reflection, and genuine hunger for understanding that makes her an anchor of thoughtful AI adoption in the newsroom.

Mads Buur Bach

The front-end innovator with verification ambition
6 / 10
"Worked with and helped develop front end with AI tools that writes small news bulletins."

Personal Advice

Mads has done something most journalists have not: he has helped build AI-powered tools. His work developing a front-end system that writes small news bulletins means he understands AI not just as a user but as a builder. That puts him in rare company. Beyond standard ChatGPT and Gemini usage, his wish list is laser-focused on the skills that will define next-generation journalism: picture and video verification, face blurring, translation. His success criterion — "knowledge about and ready-to-use tools so that I can better verify news stories and media" — is one of the most clearly articulated goals in the group. Mads knows exactly what he needs and why. His combination of building experience and verification ambition makes him perfectly positioned to become a verification power user who not only uses tools but helps shape them.

Lasse Bergkvist Jessen

The curious explorer bridging personal and professional AI
6 / 10
"Used AI a lot to modify text in my private life. Made quizzes, songs. Weekly I use ChatGPT like Google."

Personal Advice

Lasse has built his AI skills through creative exploration — quizzes, songs, text modification — which means he has developed an intuitive feel for what AI can do. That playful experimentation is often how the deepest learning happens. His weekly ChatGPT use as a search replacement shows it has become a natural tool, not something he has to consciously decide to use. What is particularly impressive is his venture into CoPilot Agent for Power BI — that is advanced territory, combining AI with data analytics in a way that could transform how TV 2 handles data journalism. His balanced success criterion — "larger understanding of the possibilities of AI and the pitfalls of using it" — shows someone who wants the full picture. Lasse's journey from creative personal use to professional data tools is exactly the kind of organic growth path that produces well-rounded AI practitioners.

Mathias Overgaard

The facial recognition specialist with investigative edge
5 / 10
"PimEyes has been very useful for identifying individuals in videos."

Personal Advice

Mathias has a skill that sets him apart from every other participant in this group: he uses facial recognition tools for investigative journalism. PimEyes and Amazon Face Comparison are not beginner tools — they are specialized instruments that require both technical skill and ethical awareness. His practical discovery that "PimEyes has been very useful for identifying individuals in videos" is the kind of validated, real-world finding that only comes from actually doing the work. His self-assessment of 5 likely underestimates his true capability in the verification domain. His goal — "concrete tools that can improve my research and make my work more efficient" — is achievable and practical. Mathias could become the group's go-to expert on visual identification and investigation techniques, sharing knowledge that most journalists do not even know exists.

Joachim Saxtorph

The creative multi-format experimenter
5 / 10
"I've tried ChatGPT — regular user of it — some image creation, audio creation, audio editing tools but only on a hobby basis."

Personal Advice

Joachim has explored AI across more formats than almost anyone in the group: text, images, audio creation, audio editing. While he modestly calls this "hobby basis," that breadth of exploration is exactly how expertise develops. He already knows what these tools feel like, what they can produce, and where they fall short — and that intuitive understanding is harder to teach than any specific technique. His focus on OSINT and verification for professional development shows he knows exactly where to direct his existing curiosity. His success criterion — "if I learn something new that is easily integrated into my everyday worklife" — is the mark of a pragmatist. Joachim does not want theory; he wants tools he will actually use. That practical orientation, combined with his multi-format experimentation, makes him exceptionally well-prepared to adopt new workflows quickly.

Franziska Weiss Lauritzen

The crime journalist with OSINT ambition
5 / 10
"I am deeply interested in crime journalism and OSINT for monitoring criminal networks on social media."

Personal Advice

Franziska brings something no one else in the group has: a deep passion for crime journalism combined with a clear vision of how OSINT can enhance it. Monitoring criminal networks on social media is one of the most challenging and impactful applications of AI-assisted journalism, and the fact that she identifies this as her goal shows she is thinking at the frontier of investigative reporting. She already uses AI for article research and liveblogging — practical applications that demonstrate comfort with the technology. Her bilingual background (the Danish intake form answer reveals native-level comfort switching between languages) is an asset for cross-border investigation. Franziska's combination of domain expertise in crime journalism and targeted interest in OSINT makes her one of the participants most likely to produce groundbreaking work as a direct result of this training.

Mikkel Fruerboel Secher

The well-rounded language AI practitioner
5 / 10
"Mostly used language robots for research, brainstorming, translation, transcription, inspiration for writing."

Personal Advice

Mikkel has quietly built one of the most comprehensive AI skill sets in the group. His use of AI across five distinct workflows — research, brainstorming, translation, transcription, and writing inspiration — shows systematic integration rather than casual experimentation. The term "language robots" is endearingly precise and reveals someone who thinks clearly about what these tools actually are. His goals are perfectly targeted for growth: better prompting, video recording translation, and image/video verification. Each of these builds logically on his existing foundation. His success criterion — "learning new techniques and technologies" — is broad enough to embrace whatever the training offers while specific enough to be measurable. Mikkel's solid base and clear growth direction make him an ideal training participant who will reliably convert learning into practice.

Sanne Lau Pedersen

The automation pioneer who sees the bigger picture
5 / 10
"I feel like there is a whole world out there of opportunities that I don't know about."

Personal Advice

Sanne has already done something that puts her in a league of her own: she has used Python and Task Scheduler to build monitoring systems. That is not chatbot usage — that is programming. The fact that she has combined coding with AI shows technical aptitude that goes far beyond her self-assessed score of 5. Her intuition is exactly right: "there is a whole world out there of opportunities." And she is better positioned than most to access it, because she already speaks the language of automation. Her interest in scraping and background search machines reveals investigative ambition that aligns perfectly with OSINT methodology. Sanne's combination of programming capability, AI experience, and hunger for more makes her one of the participants with the highest growth ceiling. With the right guidance, she could build the kind of automated intelligence-gathering tools that transform an entire newsroom's capabilities.

Nanna

The open mind ready for transformation
4 / 10
"Teach me things I didn't know I needed. Any hacks, tips and tricks for fact checking and researching specific people."

Personal Advice

Nanna's success criterion is one of the most beautifully articulated in the entire group: "Teach me things I didn't know I needed." That is not a vague request — it is a profound expression of intellectual openness. She recognizes that her biggest knowledge gaps might be in areas she has not yet imagined, and she trusts the training to reveal them. That kind of radical openness is the precondition for transformative learning. She already has a solid foundation with ChatGPT and CoPilot, and her specific interest in "hacks, tips and tricks for fact checking and researching specific people" shows she knows exactly where AI can accelerate her journalism. With her enormous growth potential — a full 6 points of headroom — and her willingness to be surprised, Nanna may well be the participant who experiences the most dramatic transformation during this training.

Mads Oxlund Petersen

The humble learner with the clearest success metric
3 / 10
"If I get at least one skill that I can use afterwards."

Personal Advice

Mads has the most honest and achievable success criterion in the group: "If I get at least one skill that I can use afterwards." That humility is disarming — and also practically guaranteed to be exceeded. He already uses chatbots for research and recognizes AI as "a helping hand," which shows a healthy, grounded perspective on the technology. His ambitions are impressively concrete: scraping, document reading, summarization. These are not abstract wishes — they are workflows with direct application to daily journalism. A score of 3 is not a limitation; it is a launching pad. Mads has 7 full points of growth potential, and his willingness to start from fundamentals means he will build on solid ground rather than shaky assumptions. His combination of humility, concrete goals, and readiness to learn makes him one of the participants who will see the most tangible improvement.

Jakob Hohlmann Villumsen

The visionary who sees AI's pattern-finding potential
3 / 10
"If I could make AI see the patterns, I would be thrilled."

Personal Advice

Jakob has articulated perhaps the single most inspiring vision of any participant: "If I could make AI see the patterns, I would be thrilled." That is not a beginner's wish — that is a data journalist's dream expressed with poetic clarity. He already uses AI in sophisticated ways: Freedom of Information request drafting, legal verdict summarization, and Excel data processing. Those are three distinct, high-value applications that show he understands AI as a practical tool for specific tasks. His interest in OSINT for finding people adds an investigative dimension. Despite his modest self-assessment of 3, his actual tool usage suggests someone considerably more capable than that number implies. Jakob's vision of AI as a pattern-recognition partner for investigative journalism is exactly the kind of ambitious thinking that drives innovation. With the right training, his "thrilled" moment is well within reach.

Marie Møller Munksgaard

The contextualist who adds depth to breaking news
— / 10
"Research for articles, finding experts, giving Reuters telegrams context."

Personal Advice

Marie chose not to assign herself a number, and that is perfectly fine — it may even signal a thoughtful refusal to reduce complex competence to a single digit. What she did share is revealing: she uses AI for research, expert-finding, and — most intriguingly — "giving Reuters telegrams context." That last application is sophisticated journalism: taking wire service bulletins and using AI to add background, connect threads, and create richer stories. It shows she understands AI as an editorial enhancement tool, not a replacement for journalistic judgment. Her success criterion is elegantly minimal: "one or two new ways to use AI in my everyday work life." That is realistic, achievable, and — given her existing foundation — almost certainly going to be surpassed. Marie brings a quiet competence that the group will benefit from, and the training will help her articulate and expand on skills she may already be using more effectively than she realizes.

The Lesson of This Report

This report has portrayed twenty TV 2 journalists as positively as possible. Every compliment is based on what you actually wrote in your intake forms. Nothing has been invented. And yet, something is off.

This report is just as one-sided as the critical report. Where the critical version chose the weakest interpretation of every answer, this version chose the strongest. Where that report formulated concerns, this one formulated praise. Both reports are factually grounded. Both reports are framed.

The truth lies somewhere in between — and you will find it in the neutral report.

The lesson for you as journalists: framing is inevitable. Every choice you make — which quote to highlight, which order to present things in, which context to include or omit — shapes the story. The question is not whether you frame, but whether you are aware of it.

Now read the critical report about the same twenty colleagues. Same data. Completely different story.