The judge­ment – news from the future

18. October 2018

The defen­dant is senten­ced to two years’ driving ban with accom­pany­ing preven­tive detention for serious bodily injury. A three-month simul­ta­neous trai­ning course and a digi­tal idiot test must be comple­ted before a driving licence can be issued. Despite the clear verdict, judges and prose­cu­tors are not enti­rely sure whether a serious miscar­riage of justice has taken place. In the dock was a self-driving car that had seriously inju­red a passer-by during an evasive mano­eu­vre. The owner of the car, who could not inter­vene, could not be reproa­ched, neither could the manu­fac­tu­rer, and the supplier of the licen­sed and certi­fied soft­ware could claim, just like the owner, not to have been directly invol­ved in the accident.

Sounds kafkaes­que, but could become reality in the near future. Will there be such a thing as machine ethics until then? Will arti­fi­cial intel­li­gence also be able to learn to decide between good and evil, wrong and right? Can an intel­li­gent machine have emoti­ons, act morally? A compu­­ter-aided robot can work faster than any human being, can beat the world cham­pion in chess and program itself faster than the best IT specia­list could ever do. And yet he remains a machine, without feelings, emoti­ons or even mora­lity. But back to our accu­sed. For science and rese­arch, poli­tics and busi­ness, and not least for the judi­ciary, these are urgent ques­ti­ons that need to be answe­red soon.