Algoritmisk etik : vad policydokument berättar om synen på artificiell intelligens

Abstract: The aim of this thesis is to look at the attitude among governments and large corporations towards algorithms in general and with a special focus on artificial intelligens. This is done by examining five policy documents produced by three governmental branches and two private companies, all of them from Europe and the US. The assumption is that policy is one of the methods that those in power can use to form norms and behavior that is in line with their agenda, therefore an examination of these policy texts could reveal the obstacles they have identified toward their goals. Depending on how these obstacles, or problems, are represented in a discourse as algorithms, a policy offers a solution. By critically analysing the problem representation instead of the solution the view of the discourse, as seen by the policy maker, is questioned and some premise that supports the problem view might fall. In doing so the representation of the problem may turn out to be different, along with the solution. The result of the analysis sugest that the governmental branches mainly discuss algorithms and artificial intelligence in terms of economy, the privacy and integrity of the people is also important as long as it does not interfere with market growth. The policys seem to rest on two premisses, the belief in artificial intelligence as a disruptive force and the promise of technology as a solution of many of mankind's problems.