DE / EN
Search
Close this search box.
DE / EN
© DeepMind auf Unsplash

Deepfake and disinformation

How the news-polygraph research project is tackling media manipulation with an AI platform

The Pope as a hip-hop mogul in a white down jacket, the violent arrest of Donald Trump or a supposed video call between former mayor Franziska Giffey and Kiev’s mayor Vladimir Klitschko: artificial intelligence is now able to suggest a false but perfectly staged reality in photos or videos that appear to be real. Although this has entertainment value, we have to ask ourselves the following question about many articles or social media posts: what is true and what is not? Because anyone can publish these days. There is no need for sources or multi-stage verification in social media. This is a challenge for journalists and media companies – they have to decide within hours whether the material they receive is newsworthy and should be included in their reporting.

How do you verify these sources? What standards are used to classify them? To combat propaganda and fake news, we need good tools and awareness-raising at all levels. The news-polygraph research project, coordinated by transfermedia at the Media Tech Hub, is currently developing such a tool set. The platform, a kind of digital lie detector, is intended to make the work of journalists easier in the future.

Checking facts is and always has been one of the central tasks of journalism. What is new is the massive amount of data – in addition to raising awareness, good, reliable tools are needed to be able to make valid statements. Many media companies have fact-checking editorial teams or offer further training. And some have put together a kind of toolbox themselves, such as a folder with various programs. These then have to be selected somewhat laboriously. news-polygraph starts with existing processes, optimizes them and brings them together in one interface. In the run-up to the project, the research alliance exchanged ideas with media partners and asked about their needs. The aim of the project is to provide a platform that offers quick, easy and intuitive assistance.

Which content is manipulated – which is not?

transfermedia COO Claudia Wolf, who is the alliance coordinator for the project, explains: “From a scientific point of view, we would like to avoid the popular term ‘fake news’ in our research project. We are interested in verifying information. The decision as to whether it is ‘fake’ or not is ultimately always made by people. We have designed a platform that provides media professionals and journalists with tools to quickly evaluate texts, video, audio and photos. This is particularly important for newsrooms.”

An intuitive user interface allows users to drag and drop in media clippings and have them checked by an AI. Models and services run invisibly in the background. The journalists only decide in advance what they want to check, determine the order of the query – and whether they want to integrate crowd panels.

While text and image verification in editorial offices is already advanced, the verification of video sequences is much more complex. This is where the Fraunhofer Institute comes into play as a partner. It specializes in audio sequences and produces forensic audio analyses for the criminal investigation sector, for example. Other alliance partners such as rbb and Deutsche Welle also contribute important focal points.

The human crowd as a truth finder

Fact-checking thanks to human crowd support is a special feature of news-polygraph. If newsrooms receive material from whistleblowers, for example, and do not have access to contact persons in the respective country who can confirm its authenticity, the integrated crowd panels become active. A mountain range, trees, vegetation at a certain time of year: local people can clearly identify images and videos without technical aids. To make this query work, volunteers are grouped into crowds in advance and qualify themselves according to their specific knowledge after a registration process (via queries). The input from the crowd comes from the scientific community, including the TU Berlin, which carried out a test phase with Ukrainian native speakers during the initial phase of the war in Ukraine. In a crowd query, the program then reports back whether a predetermined, high percentage of participants made the same statement.

Here, too, the journalists ultimately decide whether they agree with the assessment. Because, as Claudia Wolf emphasizes: “It’s up to the individual to decide.” The project title “news-polygraph” also refers to this. Like a lie detector, it is only a means to an end. It does not take over the original work of journalists, they do not want to and cannot do that: “It is in the DNA of journalists to classify and evaluate content. We merely provide simplified, fast, well-functioning and trustworthy assistance.” The intensified approach through crowd support not only allows them to react quickly, but also to go into great depth with their research.

An alliance of specialists

The semantic linking will be carried out by the transfermedia team, which has already installed an ontology for metadata in the dwerft research projectand is also responsible for project management and coordination of the ten partners. The technological part is based at Fraunhofer IDMT. Other partner companies are Ubermetrics, which specializes in social media snippets, neurocat, which checks AI models for susceptibility to errors, and the German Research Center for Artificial Intelligence and Delphai, which provide a Google alternative for B2B purposes.

In the future, the MediaTech Hub will be involved in news-polygraph on the technical/media level and mabb for legal matters. Upon successful completion of the research project, market-ready software is planned that media companies can integrate into their own systems and which will be continuously updated.

The project officially entered the implementation phase on May 3, 2023. Media professionals can already look forward to November: That’s when the project partners will present what they can expect from news-polygraph in the coming years at an event.

More about the MTH Blog

The media technologies of the future are already being used today – not only in the entertainment sector but in a wide variety of industries. For our MediaTech Hub Potsdam blog, we talk to tech enthusiasts, entrepreneurs and researchers once a month and tell the stories behind their innovative business models, ideas, projects and collaborations.