Bratislava, 15 January 2024 – Elv.ai is a Slovak startup that combines artificial intelligence with human moderators to make the internet safer for debate. It also protects the credibility of media and public institutions on their social networks. Its mission, the startup’s further activities, and European expansion will now be supported by a half-million EUR investment from the Slovak CB ESPRI Impact One fund, which focuses on projects with measurable positive social benefits. Thanks to the investment, Elv.ai has launched its product in the Czech Republic, where it protects discussions and social networks of Prima TV and the Economia publishing house. Since its launch in Slovakia, elv.ai has recorded over 2,500,000 hateful or disinformation comments out of more than 16,000,000 and has identified over 5,000 fake profiles. In the Czech Republic, out of more than 2 million monitored comments on 44 profiles, up to 504,000 were hateful content, which is 24%.
„We perceive that the media space has changed significantly and that the need to protect online discussions and social networks from hatred is growing. In our native Slovakia, we monitor and moderate the discussions of roughly one-third of the media market and help a number of major institutions that want to control the posts of their online discussions: without misinformation, hoaxes or hate speech. We are pleased that even in the Czech Republic, we have quickly found conscious media houses that consider it their duty to protect their fans and users from hateful and misinformation content,” says Jakub Šuster, co-founder and CEO of elv.ai.
Intelligence, both human and artificial
The Slovak startup was initially founded as a response to hybrid warfare and Russia’s invasion of Ukraine in 2022. It was founded as a non-profit project by the Slovak communications agency New School Communications that put together TV presenters to filter hate content. One person can manually sift through up to 300 comments in an hour. However, the first months of operation have shown that such a service is in high demand among Slovak media and institutions. That’s why, in May last year, the agency spun off the service into a separate startup that began developing its own AI tool for text analysis
Today, elv.ai relies on a combination of technology and human factors. Because of the dynamics and context of language, it uses machine learning and human moderators who control the AI and recognise possible linguistic nuances not picked up by technology. The result is the transparent removal of inappropriate and hateful content in the shortest possible time. In fact, according to internal statistics, more than 25% of comments on social media constitute hate speech, hoaxes or misinformation. Thanks to the unique combination of AI and moderators, this number can be reduced to 3-8% using elv.ai.
The four co-founders behind elv.ai are Jakub Šuster, Gabriel Tóth, Ľudovít Ďurman and Milan Dubec. All of them share an interest not only in technology but also in political and public affairs. Since the project’s launch in March 2022, elv.ai has hidden or deleted more than 2,500,000 hateful or disinformation comments out of more than 16,000,000 and has identified more than 5,000 fake profiles.
Zero tolerance for disinformation
The European expansion launched by elv.ai in the Czech Republic was supported by the CB ESPRI Impact One investment fund with EUR 500,000. The Slovak fund focuses on supporting companies and non-profit organisations with a positive social impact in protecting human rights in education, training and research of information services. By the end of 2023, it has invested up to €18.6 million from the European Social Fund and private sources.
„We decided to invest in elv.ai because we believe that cultivating the online space improves the social atmosphere as a whole. As an impact fund, we are interested in the financial return and the positive impact on society. We welcome efforts to use technology to mitigate misinformation content on social media, which should serve to connect people, not spread hate. We see the potential of the elv.ai tool not only in the Czech and Slovak markets but also elsewhere in the European market,” says Miroslav Beblavý, CEO of the CB ESPRI Impact One Fund.
Renowned Slovak media and public institutions currently use the tool. Among the largest are the television stations Markíza and JOJ, the radio stations Expres and Evropa 2, and the online magazine Refresher. elv.ai currently manages control over the content of discussions of more than 100 profiles for 30 clients, including Slovakian social media of the European Commission. New clients in the Czech Republic include Prima TV and the Economia publishing house.
Since its launch in the Czech market, elv.ai has seen more than 2,069,514 comments from the 44 profiles managed by its clients and partners in the Czech Republic. Of these, 504,506 comments were hate content, representing a 24% share. This is a 9% increase in the Czech Republic compared to the 15% average number of hidden content that elv.ai has been monitoring since its launch.
