thumbnail_logo_ai-code.png

logo_ai-code.png

The **AI-CODE** mission statement: To identify and understand synthetic content on next generation social platforms – and to provide professionals with tools and guidelines for the new media landscape they're facing.

thumbnail_logo_cleanfeed.png

logo_cleanfeed.png

**CleanFeed** is about developing tools and frameworks to verify the origin and authenticity of digital content, while also creating transparency mechanisms that help audiences understand sourcing and editorial choices.

thumbnail_logo_fakeometer_interim.png

logo_fakeometer_interim.png

**Fake-O-Meter** is building an AI-driven assistant designed to promote awareness and resilience in the face of digital disinformation.

large_logo_genki4media.png

medium_logo_genki4media.png

small_logo_genki4media.png

thumbnail_logo_genki4media.png

logo_genki4media.png

**GenKI4Media** is an R&D project that explores how generative AI can transform workflows for SMEs and news organizations across Germany.

Photo of a person holding a smart TV remote, with a TV screen in the background (by Glenn Carstens-Peters, Unsplash)

large_post_mosaic_project.jpg

medium_post_mosaic_project.jpg

small_post_mosaic_project.jpg

thumbnail_post_mosaic_project.jpg

post_mosaic_project.jpg

The **Mosaic** consortium is building Europe's first unified streaming platform for multilingual TV and video content.

small_post_news_polygraph_logo.png

thumbnail_post_news_polygraph_logo.png

post_news_polygraph_logo.png

**news-polygraph** aims to develop a multimodal verification platform that relies on artificial and human intelligence to analyze digital content.

large_logo_padse.png

medium_logo_padse.png

small_logo_padse.png

thumbnail_logo_padse.png

logo_padse.png

**PADSE** is a digital forensics project that explores algorithms and tools for analysing audio deep- and shallowfakes, with a focus on specific speakers and their linguistic characteristics.

**Bahia Albrecht** is a digital rights project manager and tech consultant. She focuses on topics like AI governance, digital security, and freedom of expression, advising projects and contributing to research and policy. As a trainer, Bahia develops programs to counter disinfo and build media literacy. With expertise in political science and education, she champions inclusive policies and ethical technology use.

medium_team_bahia_albrecht.jpg

small_team_bahia_albrecht.jpg

thumbnail_team_bahia_albrecht.jpg

team_bahia_albrecht.jpg

**Hala Attig** is an engineer focused on machine learning applications. She is particularly interested in using NLP and data analysis to make information more accessible. Before joining DW as an Innovation Manager, Hala worked as a software developer and R&D project manager for other media companies.

medium_team_hala_attig.jpg

small_team_hala_attig.jpg

thumbnail_team_hala_attig.jpg

team_hala_attig.jpg

**Ipek Baris Schlicht** is a computer scientist and innovation manager with both university and industry research experience in AI. A lot of her work is about using technology to facilitate fact-checking processes. In general, she loves to derive meaningful questions (and answers) from social data. Next to her work for DW (ReCo and AI Pro), Ipek is pursuing her Ph.D. on early detection of fake news and hatespeech in the health domain.

medium_team_ipek_baris_schlicht.jpg

small_team_ipek_baris_schlicht.jpg

thumbnail_team_ipek_baris_schlicht.jpg

team_ipek_baris_schlicht.jpg

**Julia Bayer** is an investigative journalist who uses OSINT to expose human rights abuses and environmental crimes. As Innovation manager at ReCo, she works on verification and deepfake detection projects such as PADSE, InVid and Truly Media. Julia also trains journalists in OSINT techniques and digital verification, and is the founder of Quiztime. Together with her team at ReCo, she developed Sharpen your Senses and How to Verify.

small_team_Julia Bayer.jpeg

thumbnail_team_Julia Bayer.jpeg

team_Julia Bayer.jpeg

**Ruben Bouwmeester** brings concept development skills, profound graphical expertise, and an urgent drive to get things done. He has a background in project management, concept development, and design, holding a B.A. from The Hague University. As a ReCo innovation manager, Ruben specializes in product development and verification technologies.

team_ruben_bouwmeester.jpg

**Andy Giefer** is a ReCo innovation manager in love with good design and new technology. He creates smart, artistic applications and has a background in telecommunications engineering. Before joining DW, Andy worked for the BBC’s R&D department.

medium_team_andy_giefer.jpg

small_team_andy_giefer.jpg

thumbnail_team_andy_giefer.jpg

team_andy_giefer.jpg

**Farina Hamann** is a communication and interaction designer focused on user experience design in immersive media. She enjoys supporting and connecting the XR community, hosting virtual events and giving lectures about XR. At DW she’s contributing with her experience in the fields of virtual and augmented reality.

small_team_farina_hamann.jpg

thumbnail_team_farina_hamann.jpg

team_farina_hamann.jpg

**Nico Heise** focuses on interdependencies between social media and news. At DW ReCo, he has also specialized in drones. Nico is a former media industry top executive, holds a PhD in European law, and has worked as a media lawyer for more than 15 years. He teaches as a Professor for Media Law in Potsdam.

thumbnail_team_nico_heise.jpg

team_nico_heise.jpg

**Lynn Khellaf** is a data scientist with a background in AI. They have a passion for smart applications, computer vision, neural network design, and making sense of complex data systems. In the past, Lynn has mainly worked at universities and is now happy to apply their skills in the domain of disinformation detection and other projects here at DW.

medium_team_lynn_khellaf.jpeg

small_team_lynn_khellaf.jpeg

thumbnail_team_lynn_khellaf.jpeg

team_lynn_khellaf.jpeg

**Tim Koch** has worked with the DW ReCo team since the early days. He has extensive knowledge in fundraising and the management of EC co-funded projects. Tim holds a diploma in mineralogy, which makes him the steady rock in the surf of innovation. 

thumbnail_team_tim_koch.jpg

team_tim_koch.jpg

**Eva Lopez** is an innovation manager at DW ReCO and also part of DW’s data journalism team. She’s fascinated by numbers – and a news junky, absorbing what’s going in the analogue and digital world. No wonder she decided to go for an interdisciplinary education that allowed her to dive in to the world of natural sciences, analytics, and journalism. Eva holds a Master’s degree in data journalism from Tilburg University.

thumbnail_team_eva_lopez.jpg

team_eva_lopez.jpg

**Mirko Lorenz** is a hands-on journalist with software development skills. For the last 15 years, he has worked on technologies for newsrooms. In the international DDJ community, he is best known for launching the acclaimed chart creation tool Datawrapper. Mirko is also a co-author of the Data Journalism Handbook. Before joining DW, he worked for a variety of media companies in Germany and the US.

thumbnail_team_mirko_lorenz.png

team_mirko_lorenz.png

**Kay Macquarrie** is an Innovation Manager at DW ReCo and holds a Master's degree in Multimedia Production. His work focuses on human language technologies (HLT), with a particular emphasis on overcoming linguistic and accessibility barriers through advanced technological developments. More recently, image verification tools have also become a key component of his day-to-day life.

thumbnail_team_kay_macquarrie.jpg

team_kay_macquarrie.jpg

A prolific creator of concepts and tools, **Nico Patz** has carried out research projects for ages. His focus is usually on topics like e-learning, interactive media, collaboration, and inclusive user experiences. Nico’s M.A. in English literature and History of Arts goes back to his passion for storytelling, his skills in AV production originate from more than three decades in the performing arts. Prior to his job at DW innovation, Nico worked for public service broadcaster rbb and TU Berlin.

medium_team_nico_patz.jpg

small_team_nico_patz.jpg

thumbnail_team_nico_patz.jpg

team_nico_patz.jpg

**Alexander Plaum** is an online editor, concept creator, and media consultant with a keen interest in IT, pop culture, and politics. Before joining DW ReCo as an innovation manager, Alex worked for several other organizations in the media, tech and education business. Ages ago, he also obtained an M.A. in literature studies, sociology, and political science.

thumbnail_team_alexander_plaum.jpg

team_alexander_plaum.jpg

**Axel Primavesi** is a journalist. He has worked for DW for more than a decade, putting a focus on adventurous reporting in recent years. As much as adventure, Axel loves innovation. At ReCo, he gives projects a journalistic reality check and brings his storytelling skills to the table.

medium_team_axel_primavesi.jpeg

small_team_axel_primavesi.jpeg

thumbnail_team_axel_primavesi.jpeg

team_axel_primavesi.jpeg

As the head of DW’s ReCo team, **Wilfried Runde** is responsible for the overall planning and coordination of projects and budgets – and always striving to improve journalism’s creative toolset. Wilfried worked as an information specialist, freelance print and TV journalist, and online editor before joining DW. In his leisure time, he loves to “spin a tune”.

thumbnail_team_wilfried_runde.jpg

team_wilfried_runde.jpg

**Anna Schild** is an innovation manager for DW ReCo and holds a degree in media and communication studies. She believes in the power of verification with OSINT. Anna is always eager to manage projects that translate research into practical tools: for journalists, human rights activists, and everyone else who needs to verify information.

large_team_anna_schild.jpeg

medium_team_anna_schild.jpeg

small_team_anna_schild.jpeg

thumbnail_team_anna_schild.jpeg

team_anna_schild.jpeg

**Jochen Spangenberg** is the deputy head of DW’s ReCo team. He has been involved in R&D projects since 2003. Previous to joining DW, Jochen worked for the BBC in London and had an exciting time in the new media economy. His passions are journalism, verification, disinformation detection, media literacy, and human rights issues. Jochen also lectures at Freie Universität Berlin and is a board member of EDMO, CEDMO and GADMO.

thumbnail_team_jochen_spangenberg.jpg

team_jochen_spangenberg.jpg

**Daniela Späth** is a project manager at DW Lab – and a DW-trained reporter, video journalist, and social media editor. She loves to experiment with new platforms, technologies, and code. Prior to joining DW, Daniela freelanced for SPIEGEL Online and Gruner + Jahr. Daniela is a fellow of MIZ, DAAD, FCP, and a graduate of the Erasmus Mundus Master’s in Journalism, Media and Globalisation.

thumbnail_team_daniela_spaeth.jpg

team_daniela_spaeth.jpg

**Olivia Stracke** is a journalist and innovation manager with academic training in literary studies (B.A.) and cultural anthropology (M.A.). At DW Innovation, her professional focus spans XR and immersive journalism, AI-driven media environments, and disinformation research. She has studied in Hamburg, Nantes, and London – and freelanced for a number of German media including UNI SPIEGEL, DIE ZEIT, and NDR.

small_team_olivia_stracke.png

thumbnail_team_olivia_stracke.png

team_olivia_stracke.png

**Tilman Wagner** is a DW ReCo liaison at Deutsche Welle's AI Systems and Products team. His main responsibility is that of a product manager for DW internal AI systems. He keeps exploring the deep sea of internet data on the lookout for new AI applications, models and processes as well as other precious useful (and fun) gems. With a diploma in media technology and a Master's in international media studies, he combines tech savviness with a sound knowledge of the media landscape and current trends.

medium_team_tilman_wagner.jpg

small_team_tilman_wagner.jpg

thumbnail_team_tilman_wagner.jpg

team_tilman_wagner.jpg

small_dw_innovation_key_visual.jpeg

thumbnail_dw_innovation_key_visual.jpeg

key_visual_dw_innovation.jpeg

Accessibility

AI, short for artificial intelligence, is hard to define. Some jokingly refer to it as cool stuff computers can't (or couldn't) do, some equate it with machine learning (ML) that leads to sophisticated pattern recognition, recreating, and remixing. Pretty much all experts agree AI has a lot to do with teaching software how to be autonomous (i.e. perform complex tasks without guidance at some point) and how to be adaptive (i.e. perform better by learning from experience). If something is based on a lot of (rather straight-forward) scripts and queries, it's not AI, but automation – at least by today's standards. In media tech, AI and automation (a.k.a. news algorithms) are almost everywhere now, namely in content monitoring, creation, distribution, provision, personalization, moderation, exploitation, and archiving.

large_topic_ai.png

medium_topic_ai.png

small_topic_ai.png

thumbnail_topic_ai.png

topic_ai.png

AI & Automation

A blockchain is – first and foremost – a digital directory of transactions, a kind of modern ledger that can be used for all kinds of purposes (not just for crypto currencies like Bitcoin). Individual entries/blocks are joined together like links in a chain (hence the name of the technology). Every transaction/change in the chain is checked and validated by a (substantial) number of other computers in a peer-to-peer network. The idea is to create a safe, distributed database for parties who don't know each other or don't trust each other or can't come up with an official admin – or all of the above. There are different types of blockchains, and there are also different types of distributed/digital ledgers (DLTs). The opposite of a blockchain (or a DLT) is a regular, centralized database. In media tech, blockchain tech is often used to solve IP challenges ("how can we access content in a smart way and make sure creators get paid for it?") and verification issues ("how can we prove this content is legit and has not been tampered with"?).

large_topic_blockchain_3.png

medium_topic_blockchain_3.png

small_topic_blockchain_3.png

thumbnail_topic_blockchain_3.png

topic_blockchain_3.png

Blockchain & DLT

DDJ is short for data-driven journalism. The term refers to stories that focus on numbers, statistics and visualizations thereof (e.g. in charts and infographics), but also to the very process of going through and analyzing (large) sets of data – usually with the help of digital tools – in order to create such content. DDJ, which can be closely linked to data science, has continuously evolved and now includes all kinds of sources, ranging from "traditional" research on the web to vast digital file dumps to internet of things (IoT) sensor recordings or satellite images and earth observation (EO) data.

large_topic_ddj.png

medium_topic_ddj.png

small_topic_ddj.png

thumbnail_topic_ddj.png

topic_ddj.png

HLT stands for human language technology and is basically a mix of applied linguistics and applied computer science. HLT people use computers to aggregate, analyze, produce, alter, or react to human text and speech. This is also known as NLProc or NLP (natural language processing), which is then combined with computational linguistics (CL) and app development. HLT has become a pillar of digital media and makes prominent use of AI approaches. There would be no automatic transcription, translation, subtitling, and voice-over generation without HLT – and no chat bots, voice assistants, meta data generators, and summarization tools.

large_topic_hlt.png

medium_topic_hlt.png

small_topic_hlt.png

thumbnail_topic_hlt.png

topic_hlt.png

In journalism and media, verification refers to a number of different, but related processes concerned with finding the truth ("veritas" in Latin). It's about investigating the facts/details of a story (fact-checking), examining pieces of content with regard to their validity, and establishing if the narrative holds up. Nowadays, journalistic verification is usually done with the help of search engines, digital maps, and lots of other tools. It has become increasingly complex, and there's a fair amount of overlap with other disciplines, e.g. digital forensics or the broader field of OSINT (open source intelligence, which refers to publicly available sources of information, not open source software).

large_topic_verification_2.png

medium_topic_verification_2.png

small_topic_verification_2.png

thumbnail_topic_verification_2.png

topic_verification_2.png

Verification

This is a very broad term, but also one that's easy to explain. Visual journalism encompasses all methods and techniques of novel, digital storytelling that heavily rely on (powerful) images and symbols. Examples range from avantgarde drone cinematography to elaborate scrollytelling to gaudy vertical video.

large_topic_visual journalism.png

medium_topic_visual journalism.png

small_topic_visual journalism.png

thumbnail_topic_visual journalism.png

topic_visual journalism.png

Visual Journalism

XR is a hypernym that resolves to "extended reality" and includes the categories of AR (augmented reality), MR (mixed reality), and VR (virtual reality). These technologies, which often overlap, aim to do one or several of the following things: add information on top of the real world, blend objects into the real world, create a virtual world and transport viewers into it, replace the real world. The general idea is to provide content that is (a lot) more immersive than regular AV content and takes viewers to an exceptional place and/or puts them in the shoes of somebody else, thus fostering a particularly engaging and absorbing experience. When XR is used to tell a story, provide insights, and explain the world, it becomes immersive journalism. Increased levels of immersion can also be achieved via 360° degree videos, sophisticated 3D models (viewed in 1st person mode), and binaural sounds.

large_topic_xr.png

medium_topic_xr.png

small_topic_xr.png

thumbnail_topic_xr.png

topic_xr.png

XR & Immersive Journalism

This category collects blog posts that look at media tech R&D from a meta perspective. Questions discussed include: How do you run a project? How do you run a team? What do modern ideation, concept creation, and management tools look like? It's a smorgasbord of tips, tricks, and insights.

large_topic_best practice_3.png

medium_topic_best practice_3.png

small_topic_best practice_3.png

thumbnail_topic_best practice_3.png

topic_best practice_3.png

Best Practice

Interview

To provide you with relevant content we use essential cookies to monitor the usage of this website. To display external content, you need to accept all cookies. You can always adjust your settings via the [Privacy Statement](https://innovation.dw.com/privacy-statement).

DW Innovation = DW ReCo + DW Lab. We explore emerging technologies and their impact on journalism, media, and audiences.

DW Innovation

Check out this post I've come across on the @dw_innovation blog: {{link}}

Check out this post I've come across on the @dw-innovation blog: {{link}}

Check out this post I've come across on the Deutsche Welle Innovation blog: {{link}}

An image showing a highlighted geospatial information pattern in Cairo (blue tint).

large_post_spot_v2_key_cairo.png

medium_post_spot_v2_key_cairo.png

small_post_spot_v2_key_cairo.png

thumbnail_post_spot_v2_key_cairo.png

post_spot_v2_key_cairo.png

Ruben Bouwmeester introduces the new, improved version of our essential verification service.

Geolocation Tool SPOT: Reworked & Out of Beta 

Ruben Bouwmeester

In case you're not yet familiar with SPOT yet: It’s a free tool that helps investigators, journalists, and open-source researchers verify locations using natural language prompts. Just describe what you’re looking for – say, a place shown in a photo or video - and SPOT searches OpenStreetMap (OSM) for matching locations. 

For example, you could ask: 

"Find a church within 100 meters of a building with 10 stories or more and water in the City of London." 

SPOT then searches OSM data and displays all matching locations on an interactive map:

An image showing a highlighted geospatial information pattern in London.

large_post_post_v2_example_london.png

medium_post_post_v2_example_london.png

small_post_post_v2_example_london.png

thumbnail_post_post_v2_example_london.png

post_post_v2_example_london.png

## A Major Upgrade Behind the Scenes  

After two years of training and fine-tuning our own AI models, we've significantly upgraded how SPOT works. 

Our benchmarking system revealed something surprising: a larger open-weights model without custom training consistently outperformed our smaller fine-tuned model. Thanks to recent advances in AI and the availability of new models through the DW LLM Hub, we were able to make the switch. 

As a result, SPOT is now more accurate, more robust, and easier to maintain. 

The change also reduces the time and resources previously needed for generating training data, maintaining benchmark datasets, retraining models, and hosting custom infrastructure. 

Building datasets, training models, and evaluating performance gave us deep insight into how AI systems behave in real-world applications. That experience helped us build a stronger, more sustainable foundation for SPOT. 

## SPOT 2.0 Improvements 
The new version of SPOT brings several important improvements: 

-   better overall performance and reliability 
-   faster implementation of fixes and enhancements through prompt engineering 
-   easier migration to newer and more capable AI models 
-   expanded coverage of OSM tags 
-   improved search engine 
-   better recognition of entities with specific attributes (e.g. "Italian restaurant" or "vegan café")
-  improved [tutorials](https://www.findthatspot.io/tutorial/instructions) 

## New Feature: Tag Bundle Editor 

SPOT now also includes a Tag Bundle Editor, a feature that allows you to explore how semantically related OSM tags are grouped together. 

For example, you can see [how tags associated with "airport" are separated from tags associated with "military airport."](https://tags.findthatspot.io/?q=airport) 

The Tag Bundle Editor also lets you suggest changes when tag groupings don't behave as expected, helping us continuously improve SPOT's search results. 

## Free and Open Source 

SPOT is available free of charge at [https://findthatspot.io](https://findthatspot.io/). 

The entire project is open source and hosted on [GitHub](https://github.com/dw-innovation/SPOT). 

We welcome feedback, bug reports, feature suggestions, and contributions from the community. Get in touch via hey[at]findthatspot.io – or connect with us on GitHub. We'd love to hear from you.