Defudger works to advance SDG's 9,16 and 17: they use cutting-edge technology like ML to detect and stop fake news and deepfakes to counter the huge challenges they are posing to our political systems. Their technology is the means to building more transparent, equitable systems that make use of more inclusive algorithms, and this supports the base of a sustainable society.
Mónica López Calero
José María Panadero Meseguer
Universidad de Navarra
Isabel Rodriguez Tejedo
Defudger is the only solution that combines three-layer verification with state-of-the-art technologies and digital forensics techniques. It not only considers the present stage of the content, but also looks at the past using reverse image search. Additionally, they are building a solution to track the different versions of an image since its creation much like in a supply chain, using blockchain. This will allow them to determine the source of the content with certainty, opening a new era of validated content flow.
Kristof, one of the three cofounders, used to have a company which was related to the media industry. He became acutely aware of the threat of fake information and its harmful effects. Zoltan and Dominik already knew about deepfakes from their technological background and out of curiosity deep dived into the topic. This happened in late 2018, which is one year after the emerge of deepfake technology which was developing rapidly, both in terms of technological sophistication and societal impact. This made them come up with the foundations of Defudger and apply to the prestigious Idea Competition of Denmark, where they ended up as finalists.
Defudger initially had an alpha testing during the Summer of 2019 with a Sanish digital platform called Indieframe. Over the period, the initial Defudger detection algorithm filtered out 900 images, 300% more than the initial 300, and in less than 15% of the time. In the Autumn of 2019, Defudger joined APX (Axel Springer x Porsche Accelerator) in Berlin, one of the most prestigious startup accelerators of Europe. By the end of the 100-day program, they launched their beta version with Der Spiegel (one of the largest newspapers in Europe) and Deutsche Presse Agentur (largest content distributor in Germany). Defudger is now in talks with CNN, Sky, Bertelsmann and Facebook about potential future testing partners. They are planning their fully commercial release by the end of Spring 2020. Now the company employs 8 people who are determined to cleanse the internet from misinformation. The innovation has tremendous potential in economic terms. A report by CHEQ and the University of Baltimore estimated the annual cost of fake news in about 78 billion dollars for the global economy.
Their primary short-term goal is to raise awareness about the dangers of the fact that now anyone who has access to a powerful computer is able to create very high quality manipulated content. By doing this, media houses will create their own fact-checking units which will make sure that the content they provide is valid. Defudger will help to automate the process of being able to validate vast amount of digital content. In the long-term their true purpose is to protect our society and democracy from misinformation. Fake news spreads like a virus in our modern world and they want to be the antidote against it. They see the digital forensics industry becoming similar to the anti-virus industry, where corporations are going to use multiple solutions against the threat of fake news. Defudger wants to lead this movement and stay the state-of-the-art solution in the long run.
The innovation is the core of the business itself. Through its innovative use of deep learning techniques, Defudger is placing itself as a leader in this very new industry. They have created a new service that offers advanced solutions to fight fake news. This will be of tremendous importance for firms wishing to protect their image, and for news outlets or any other service that wants to provide reliable information to its users. Ultimately, by helping combat fake information defudger will contribute to a better society
Money wise, Defudger follows a SaaS business model with a yearly recurring fee. According to their beta testing period, their partners have expressed the willingness to pay between €5,000 - 7,500 per year/license in license fees for the technology. Their Serviceable Obtainable Market (SOM) comprises of the digital forensics’ tools market in Europe and in the US. The potential to reach of around €8.4B in 2027. Within the expected SOM: in their 7-year financial projection they expect a market share of 0.2%. This means that, after the prototype trial, they are updating their growth strategy and incl. revenues from consulting services and migration of existing databases to achieve a 1% market share (€84M). Currently, Defudger is not bringing in profits; they have testing partners and a soft money income.
The company has also expanded its staff, growing from the inital three cofounders to its current team of eight employees.
Current and past testing partners were astounded by the technology Defudger provides. By saving time for the journalists, they are able to do what they are the best at: writing articles. Moreover, by increasing each company's content validation capabilities, their consumers had more trust and willingness to obtain information from them. The true social impact with Defudger’s technology is trust and the feeling of security. Without fact-checking and automated detection they could reach a state when seeing is no longer believing. Eventually, when they lose trust in each other, they end up breaking up as a society. This is one of the greatest threats to democracy today.
Dominik Kovacs, Chief Operating Officer
Keep this story going! Share below!
Copenhagen, Copenhagen, DK
Business Website: https://defudger.com/
Year Founded: 2018
Number of Employees: 2 to 10
Defudger is a complex, three-layer detection tool for the verification of digital content. The tool reveals forgery and falsification of images, videos and deepfakes with the highest accuracy of any detection tool.