Data science may soon expose fake news

Nikolaj Tatti, a recently appointed associate professor in data science, believes that in the future, businesses specialised in information security will have an even greater need for the expertise of data scientists.

For laypeople, information security primarily means antivirus programmes used on computers, but Nikolaj Tatti asserts that it is much more.

In February 2019, he began working as associate professor of privacy-aware and secure data science at HiDATA, the Helsinki Centre for Data Science.

“Information security is such a broad concept. In its simplest form, it indeed denotes antivirus and anti-spam software, but at the other extreme, you have complicated attacks targeted at large organisations. The latter end up in the news, for example, when data concerning large user groups have been stolen,” Tatti explains.

He believes that information security will still keep data scientists busy in the future. Viruses and attacks keep changing all the time, making even artificial intelligence unable to identify the very latest versions. This is where humans are needed.

“What fascinates me in information security is how much it resembles a biological process. Constantly evolving viruses are not called viruses for nothing. Figuring out the underlying biological process is extremely laborious. In information security, getting to the root of the problem is slightly easier. You can actually cross the finishing line,” Tatti describes.

Lately, information security businesses have increasingly started to utilise the expertise of data scientists. After all, keeping information secure requires the processing of enormous masses of data.

“Data mining is the trade of data scientists. It’s very burdensome to go through online files only visually. It’s here where we can help information security specialists by developing tools with which to uncover certain signals,” says Tatti.

Tatti himself aims to focus on modelling viruses and attacks at the University. In other words, on the common features of different attacks and their progress.

“My work is basic research, still several steps removed from applications that can be used, for example, by companies,” he points out.

Visiting the business world

Before ending up at the University, Tatti worked for a year and a half at the cyber security company F-Secure as a data science specialist, investigating how data mining could improve security products marketed to companies.

“My particular focus was on how the prevention of attacks targeted at individual businesses could be made easier. In such attacks, malicious individuals try to break into the data system of an organisation with the intent of causing harm. It’s precisely the detection of attacks carried out by individuals that’s difficult, as they are often very complicated and require the processing of large swathes of data,” Tatti explains.

Before F-Secure, Tatti worked as a researcher at Aalto University. He found his visit to the business world interesting.

“Every company is an entity of its own, conducting its business in a way slightly different from others. The work of an associate professor, on the other hand, remains largely the same regardless of the university,” he notes.

Business collaboration

At the University, Tatti is also teaching students of data science in addition to conducting research. As a teacher, he puts emphasis on independent thinking.

“Learning can be superficial, based on swotting up on rules, as in ‘tool Y is used to solve problem X’. Instead, it would be better to profoundly internalise facts and perceive their interconnections. To pause and ask why something has been done the way it has been done. It takes more work, but the effort will be repaid,” Tatti says.

In the sphere of data science, he would like to see even more cooperation between the University and businesses, believing this reciprocity would benefit all parties. Businesses would gain closer contact with research, getting to know up-and-coming researchers, while students would gain work experience and traineeship positions. And the University’s understanding of business life and the issues most concerning businesses would improve.

Intercepting fake news

Right now, the hot topic in data science and information security is also the focus of Tatti's research: targeted attacks, which have steadily become more common.

Tatti believes that in a few years, data scientists specialised in information security may be working on problems that are quite different from today.

“The identification of fake news promulgated on social media is a serious problem also encroaching on our work. It’s extremely difficult to develop a system that would be able to unquestionably recognise lies,” Tatti estimates.

As technology advances, fake videos into which the face of a person not actually appearing in it is later embedded may become a problem with dramatic consequences. So far, the related technology is clumsy, but already in a couple of years it may be impossible to say for certain visually whether the person speaking on a video really is, say, Vladimir Putin.

“Fake videos can cause a lot of harm, which makes it necessary for us to prepare to identify them,” Tatti says.