Legal issues of Artificial Intelligence. The Chess Game as a metaphor of the controversial relationship between the Human and the Machine

Nowadays, politicians are increasingly using social media and new communication techniques to attract the attention of millions of voters, preferring very direct forms of expression, such as tweets, keywords and hashtags. Clearly, the most relevant national issues are not explored because what interest the most is the largest number of likes, which, in the polls, will be translated into votes. In this context, Machine Learning algorithms are fundamental in terms of both analysing the large amount of data on the network and doing a political screening of the users.

Therefore, this makes possible to understand and predict the intentions of voters and, as a consequence, to plan election campaigns. Politicians are supported by specific companies that, by using data mining techniques, are able to identify segments of the electorate by profiling them on the net, as if they were doing market analysis. According to a study of the Department of Psychology of the Cambridge University, with hundred and fifty likes on Facebook, a Machine Learning algorithm is able to know more things about us than our family knows and with just seventy likes is able to know our personality better than our friends and roommates; just our partner seems to know more about us than the computer.

Moreover thanks to chatbots, politicians are able to customise their communicative approach in order to influence decisions of individual users. For example, in the United Kingdom, a form of computational propaganda has already taken place during the Brexit referendum. These are fake profiles, mainly created on Twitter, managed by software, which simulate to be a real person with the aim of increasing the visibility of a product and a political campaign, or to counter an opponent. As we have seen, the search algorithms, the dominance of single search engines and social media platforms clearly influence citizens’ political choices. This manipulation, in order to be difficult to be discovered, takes the form of the so-called “Filter Bubble”, that is the effect of the system of personalisation of the results of searches on websites that record the history of the user’s behaviour as for example are doing Google and Facebook. The latter are in fact able to use information such as position, previous clicks and past searches to selectively select, among all the answers, those that the users will want to see. The effect is to exclude the user from information that is at odds with his point of view, isolating him in his cultural or ideological bubble. On a large scale and at political level, thanks to the filter bubble, there is a serious possibility of conditioning slowly and gradually the public opinion for example by feeding the hostilities towards ethnic minorities and migrants and address the vote of the electors in a certain way. This issue, despite being a current theme, is starting to be anachronistic.

As citizens become increasingly frustrated with human politicians it may not be far away a scenario in which intelligent machines will run for an office. In this respect we should be aware that an AI-robot -named Alice- created by the Russian web giant “Yandex”, was initially nominated for the presidential elections in Russia, until the moment was silenced by President Putin. Moreover, in 2018, a machine named “Michihito Matsuda” placed third with a four thousand votes in a mayoral’s race in Tama City in Japan with the slogan: “Artificial intelligence will change Tama City”. Alongside Alice and “Michihito Matsuda”, we can talk about “Sam”, considered to be the world’s first virtual politician, who according to its creator Nick Gerritsen, will may be able to run in New Zealand’s 2020 general election.