The fundamental problem with data mining is that as users we are rarely aware of it and we often have no idea how the data is handled further, or who has gained access to it. For social network operators, data is the most valuable item they can trade – this is how they make their living. Our ignorance allows them to sell our data to companies that use it to phrase their offer in such a way as to draw our interest.

The usual disadvantages related to the collection of personal data are intentional and unintentional data leakage and the use of data for various dishonest activities – from sending spam to blackmailing. Apart from this, there is also the fact that social networks monitor the behavior of users even when they are using other websites – that is, browsing on the Internet.

Based on this behaviour, they then build tailor-made marketing models, which can be misused in politics, especially in a pre-election duel. This was also the case of the aforementioned Cambridge Analytica, to which Facebook was sending the information and data of its users. Cambridge Analytica was then able to target Facebook users in ways that are by the professional public considered controversial, to say the least. Examples include the Pro-Brexit campaign in the UK and Donald Trump's 2016 US presidential campaign.

Deepfake is a technology which can exchange one person’s face for someone else’s using artificial intelligence, most commonly neural networks. As a rule, they use faces of celebrities because artificial intelligence needs a lot of photos or videos to begin with – and those of celebrities are all over the Internet. This phenomenon has been the centre of attention since 2017, when the first pornographic and later political deepfake videos began to emerge. Since then, there has been an on-going debate about how harmful this phenomenon actually is.

While it might seem that it was modern technology that brought about the ability to alter video content, the opposite is true. Video creators have been manipulating the content virtually from the moment it became possible to capture reality on the film strip (leaving aside the photos that are spreading like wildfire – see, for example, President Trump’s “diet”). We are not necessarily talking about trick sequences in old films, even though truly magical (such as A Trip to the Moon by George Méliès, the father of special video effects).

Around the end of the 1930s, artificial content appeared even in news films that were shown at cinemas. Because of the lack or unavailability of the image material, a substantial part of the film had to be made in the studio. After all, it does not have to be a manipulated video, just content taken out of its original context. Most of the time, we want to see what we have long believed, and so even a slight nudge when interpreted in the right direction can lead to manipulation.

In the Czech Republic, we encountered a striking example of a manipulative video in 2007, when the Ztohoven (Out of shit) group hacked a broadcast of Czech Television with a video of a nuclear explosion in the Giant Mountains. This never happened, of course, but the footage still provoked at least outrage, if not outright panic.

If videos have been manipulated for so long, why be afraid of deepfake now? The problem is that the possibilities of producing deepfake videos have quite expanded, and the reach of such videos can be enormous. We are not talking about pornography, but about politics. It is possible to create a video in which a politician can be saying practically anything. Or (s)he does not even have to be speaking at all. (S)he may just appear in an unexpected context. In theory, such a video could cause major problems. Just imagine a deepfake video with, for example, the American president insulting the Russian one. Can we protect ourselves? We cannot rely on technology just yet. Although there are attempts at writing algorithms revealing deepfake videos, the creators of these videos are still one step ahead. This is because algorithms learn from other algorithms. It is a never-ending cat-and-mouse game.

The very best defense strategy is to accept that not everything we see in the media is actually true. We must be even more distrustful of the Internet. We have to use critical thinking and approach any media content we consume with caution. Caution must be exercised all the more when shocking content comes from an unknown or unverified source.