Don't be fooled by manipulations.
Detect them with Verifee.
Are articles manipulating me? Can I trust what I read? If you are asking yourself these questions, you are in the right place. Verifee is a tool designed to help you answer them, and what's more – it can prevent them.
Trust, but verify.
It would be an exaggeration to say that nothing can be trusted on the Internet. But the fact is that you probably regularly come across articles that try to deceive you. We understand that checking the credibility of all the articles you read is beyond your power. That's why we developed Verifee, which uses artificial intelligence to determine it for you.
Recognize untrustworthy articles in time.
Verifee is a program that helps you protect yourself from manipulations. It evaluates the credibility of the articles you read. It works on the principle of artificial intelligence, which, unlike humans, is not biased towards the sources and topics of articles. In addition, it can process a large amount of data. Therefore, it is able to learn to recognize manipulative elements in the text and safely mark them.
It can be very difficult to recognize manipulations. At the same time, recognizing them is essential in order to be safe on the Internet. Verifee looks for manipulative techniques in all the news articles you read every day. It creates a detailed analysis and assigns a score to each of them. You can then make an informed decision whether to read the article or not and whether to believe it.
Verifee shows you detailed results of each analysis. You can thus not only check it, but also find out what tactics the manipulators are trying on you. In addition to listing the flawed aspects of the article itself, Verifee will also give you specific examples of errors in the article so you can get a clear understanding of what Verifee sees as the problem.
Time to try Verifee. Analyse an article.
Project of avid students.
Verifee is a project of three students interested in artificial intelligence. They first met in 2020, when Euro weekly ranked them among the top 20 most talented Czechs under 20. A year later, with the project that laid the foundations for Verifee, they were the overall winners of Robothon 2021, the biggest AI hackathon in the Czech Republic. That same year, they partnered with Dataclair.ai, a center for artificial intelligence, to bring Verifee to a successful release. Currently, the project is run in partnership with CEDMO and Charles University, who ensure further development of the application.
Our team
featured in news
Answers to FAQs
Verifee recognizes the features that untrusted articles have in common. These may include argumentative fallacies, clickbait headlines, unattributed authors, or offensive language. Based on their frequency in the text, it determines a so-called trustworthiness score and places the article into one of four categories that quickly tell you whether you can trust the content of the article.
The credibility score is the main output of each article's analysis. As a whole it speaks about whether the information contained in the text is presented objectively, properly substantiated, and the article does not try to manipulate its readers. Its value always ranges from 0 to 100, where 0 means the lowest and 100 the highest trustworthiness.
No. Verifee does not in any way hide the articles it assigns a low score. These are accessible to the users in their original form, the application only alerts them. The readers themselves decide whether to take the score into account.
Determining the truth is a very complex problem, because it changes quickly and we sometimes cannot be sure of the truthfulness of new facts. For this reason, we focus on determining the credibility of an article. When someone is trying to manipulate you or hide something from you, they put their sentences together in a certain way. Our tool has learned to recognize these sentence structures and is thus able to analyse articles and assign them a corresponding credibility score. So we are not determining what is true, but whether you can believe what the article tells you.
No. The application does not perceive the topics of the articles as such, but the way in which they are presented. The entire system was trained on data collected in collaboration with Charles University, where the overall credibility was derived according to our comprehensive annotation methodology. The students thus evaluated the articles according to several dozen very detailed aspects of the submission form, without focusing on their topics. Nevertheless, we intensively tested the system for the presence of tendencies towards different political opinions or directions and we did not detect bias towards topics.
Verifee does not check the truthfulness of the claims contained in the article, but analyzes the credibility of the article and whether the information presented in it is balanced, quoted or relevantly substantiated and does not contain any manipulative techniques. This process does not differ in any way according to topics, politicians, or mentioned companies, so our partners cannot manipulate the results.
Supported by
