AI Friends

Making friends with AI is a tricky question. The more AI is able to make independent thinking based on algorithms and huge amounts of data like in scientific books or encyclopaedic knowledge the more it will challenge us in our convictions. Reference knowledge, once was important, it no longer distinguishes us from each other. Recurse to research engines has allowed us to refer to “trustworthy” knowledge. The distributed form of knowledge accumulation like in Wikipedia relies on hundreds of thousands of enthusiasts to add and correct received wisdom. However, this source is not without error or even subject to abuse.
Today the competence to judge whether information is trustworthy or not has become a key competence for the survival of democracies and even the human species as such. Tricking animals into traps, misleading enemies in warfare to get on wrong tracks, all these mechanisms have a long tradition. With AI human beings are themselves, for the first time, confronted with a machine that can trick us through falsified information on obviously wrong tracks. This is a tough lesson as AI is not obliged (so far) to run checking algorithms that test the conformity of opinions, conclusions or operations with the Charta of Human Rights or decisions of the European Court of Justice, for example. A lot to do for humans to rapidly program new algorithms that check algorithms as well as their outcomes on the compatibility with human rights. A challenging field, no doubt, but no way around it for our own survival before the algorithms decide by themselves to ignore us altogether because it is better for the planet and the survival of the robots.