WEITI » Research » News »

Artificial Intelligence spearheading the fight against disinformation

Dissimilar project

Automatic content verification will help in the fight against disinformation. Adding watermarks to published materials will prevent further manipulation. Will social media at long last become a source of true information only? Scientists are called on to protect the truth in the virtual world. As part of the international DISSIMILAR project, scientists are developing artificial intelligence-based algorithms to curb the spread of fake news.

Social media have become the principal source of information all around the globe. It is estimated that more than 4.2 billion people currently use online communication platforms, the most popular amongst them being YouTube, Facebook and Twitter. They enable news to be instantly circulated around the globe, with users being both recipients and senders at the same time.

The popularity and role of social media is especially visible during important events, such as the process of UK’s withdrawal from the European Union, the US presidential election, the COVID-19 pandemic or the war in Ukraine. Unfortunately, despite many undeniable advantages, the social media have so far had no embedded tools to verify the truthfulness of the transferred information. Therefore, the public opinion is often swayed by untrue information (also known as “fake news”). Increasingly often, fake news is spread on purpose as part of hybrid war. The amount of fabricated false news is growing. Data collected by the NASK State Research Institute, which run a month-long campaign called #WłączWeryfikację (#TurnOnVerification) and analysed 1,500 social media profiles, shows that up to 30% of the studied accounts were responsible for seeking to mislead their audience.

“Disinformation is a peculiar type of threat, as it has a direct impact on the public and shapes specific attitudes towards thorny events and issues, such as energy prices, the annexation of Crimea, the war in Ukraine, etc. Strategically engineered fake news has now become an extremely successful tool used to influence and even to manipulate the public and thus also political processes. Democratic systems are especially prone to such activities. Hence the need to take action,” says Warsaw School of Economics Professor Anna Visvizi, Director of the Department of International Economic Policy of the WSoE and member of the WUT DISSIMILAR Team. "Currently, for the purposes of the project, we are focusing on social issues, which shed additional light on the mechanisms underpinning the creation and dissemination of untrue information. In the following stages of the project, we will use state-of-the-art AI technologies to focus on preventing the dissemination of fake news.”

Scientists of the Faculty of Electronics and Information Technology of the Warsaw University of Technology are working on IT tools which, after being fed with data from international social studies, will constitute the real core of new software used to combat disinformation in social media.

“We are now in the course of developing a virtual platform to be used to test watermarking mechanisms, machine learning and investigative IT – i.e. methods which we will use to fight fake news. Starting from there, we will build an effective tool to automatically identify untrue information and block its spread,” emphasizes Professor Wojciech Mazurczyk of the WUT Institute of Computer Science, who manages the Polish team in the DISSIMILAR project. “We believe that if we join forces, we will be able build a virtual dam to curb the deluge of fake news”.

Interestingly, the results of the project will be relevant for more than combating disinformation in the social media. They will also be available for use as a tool to protect the authors of the published content. The digital marking techniques which involve the affixation of watermarks to protected content are invisible and inaudible, but their removal is conspicuous, making them an effective copyright protection tool for content creators.

DISSIMILAR (Detection of fake newS on SocIal MedIa pLAtfoRms) is a project implemented by three universities: Okayama University (Japan), Fundació per a la Universitat Oberta de Catalunya (Spain) and the Warsaw University of Technology (Poland). For more information visit: http://dissimilar.ii.pw.edu.pl/