Page - 152 - in Limina - Grazer theologische Perspektiven, Volume 3:2

152 | www.limina-graz.eu Frank G. Bosman | The turning of Turing’s tables To test an A.I.: Alan Turing’s test 9S’s interpretation of the copulating machine men was that they “don’t have any feelings” and “just imitate human speech”. This idea of emulat- ing human and artificial behaviour and inhibition – both of which are no- toriously bad at doing so (Pennachin/Goertzel 2007, 8) – is the focal point of the famous Turing test, named after its creator Alan Turing (1912–1954), and its reversed version. The original test was targeted at an artificial intel- ligence’s ability to be indistinguishable from a human interactor as judged by another human (Turing 1950). And although the test has been criticized, for example in the “Chinese Room” thought experiment (Searle 1980), it is still a very important and decisive moment in the short but already boom- ing history of the development of artificial intelligence (Moor 2012). The test also has a reversed version where the role of the judge is given to an A.I. instead of to a human participant (Schieber 2004, 13). These kinds of tests are frequently used in web applications, with the most famously called CAPTCHA (Ahn et al. 2003), giving only human users access to cer- tain features. The A.I. in charge of the process has to be able to reliably tell which user is human and which is artificial. Since, as has already been said, both human and artificial entities are notoriously bad at emulating one an- other, this reverse test is usually reliable, although not always (Crockett 1994), especially concerning usability (Brodić/Amelio 2019, 24–25). The original Turing test, although no longer used very much in its original context of artificial intelligence research, has established itself very well in the world of modern fiction, for example in films like Blade Runner (1982) and Ex Machina (2014), or games like Bioshock 2 (2007) and Metal Gear Solid: Peace Walker (2010). And within these narrative-fictional contexts, the traditional Turing test changes shape, morphing into an anthropological thought experiment. Thought experiments are “experiments” that exclu- sively take place on a cognitive-imaginative level, usually because more “classic” empirical ones are not (yet) possible (Zeimbeikis 2011). Famous examples include Schrödinger’s cat on the Copenhagen interpretation of quantum mechanics (Schrödinger 1935), and Maxwell’s demon on the sec- ond law of thermodynamics (1872). In fictional contexts, the Turing test is narratively used, as a thought experiment, to reflect on the fundamental anthropological question: what does it mean to be human? They “don’t have any feelings” and “just imitate human speech”.