Browsing the internet
Photo by Vojtech Okenka
As part of what is being called the third wave of artificial intelligence (AI), two Polish entrepreneurs have developed software that can search for and detect fake news, hate speech and violence across online platforms.
Michał Wroczyńslo and Gniewosz Leliwą developed the technology through their company, Samurai Labs. The precision of the software is a first for the industry and is providing hope and a potentially very powerful weapon in the fight against the tidal wave of online disinformation and internet trolls.
The entrepreneurs gave an interview with Polish-language website NaTemat in which they explained that second wave AI software uses algorithms that were developed from a large data set. In practice, this means that is is restricted to finding phrases that already exists within a database.
What Mr Wroczyńslo and Mr Leliwą have pioneered at Samurai Labs is software that has the ability to deduce and alter the algorithm accumulatively, which results in much greater precision.
“Our system is no longer sensitive only to keywords, such as cursing, but is able to observe language phenomena such as negation, dependent speech or references to the first person, and capture offensive messages within them,” said Mr Leliwą.
The greater detail provided by the software allows police and authorities to tackle various problems more efficiently as they will already know what is happening in more depth and with more nuance.
Samurai Labs has already been used in the context of online sexual predators and pornography, in particular with regards to paedophilia, but has also been used in the fight against online hate speech and fake news.
“It should be said how important it is for Samurai to distinguish these different types of violence, because the fact that we detect violence does not necessarily allow us to react to it. Personal attacks, threats to life and paedophilia are different. Each type of violence has a different level of threat and requires a different action,” Mr Leliwą told NaTemat.
The start-up was recently hired by UK police to work alongside Cardiff University's Hatelab in an investigation looking at hate speech against Polish people on British social media in the run up to the Brexit referendum in 2016.
With over 900,000 Poles living in the United Kingdom, they are now the largest national minority in the country. An initial pilot study by Samurai Labs showed that around 5% of material published on UK social media about Polish people had a negative or offensive connotation.
Following the result of the Brexit referendum, police recorded a spike in hate crimes; an increase of 57% on the previous year. This trend continued and in 2017/18, there was a further increase of 17% on the year before.
Mr Wroczyńslo stresses that this kind of software is imperative in an age of increased cyber warfare, citing the threat posed by a flood of disinformation and potential electoral interference by Russia and other international players.
He believes that this technology could be used to actively counteract this type of cyber warfare as it could accelerate the response by Facebook and other social media platforms to real time, as opposed to the current 24 hour time period.
This would result in sufficiently limited air time given to fake news and hate speech, thus limited the real world impact.
Last year, Samural Labs received investment from Robert Lewandowski, the captain of the Polish national football team.
Mr Lewandowski, who currently plays for Germany's Bayern Munich said at the time: “I am glad that technology is being developed in Poland which will help people from around the world in the fight against the dangerous phenomenon of cyberbullying.”
Back to Homepage
Back to Technology & Innovation