https://www.heise.de/newsticker/meldung ... 61722.html(German language link)
Heiko Maas, SPD Justice Minister, has spoken out against computer algorithms and AI, which could discriminate against people. He says that evolution in technology should not lead to a devolution in society.
He wants to bring out an anti-discrimination law, which will ensure that algorithms and programmers do not make decisions about parts of the population. He calls it an advantage-free (direct translation) decision making process.
In this, he is talking about boosting prices for products based on geolocation or type of device (E.g. the known cases on services like Amazon, where Apple users are charged more than Windows and Android users for the same products), our emotional state based on how we type, or more dramatically, increased (health) insurance costs based on search behaviour etc. It would also be used to reduce the impact of echo-chambers and filter bubbles in social networks.
Other examples include: In the USA the pre-sorting of job applications through self-learning algorithms, which the US Justice Department fears will lead to criminals falling back into criminal ways as they are pre-ostracized from the working community. In China they are working on a "Citizen Score" to calculate the social reliability of citizens, which is linked to hard sanctions. And in Australia, the Treasury is using a routine to identify and calculate back taxes on millions of people.
"If we reduce people to their past, we can obstruct their future chances," warned Maas. With justice and the police, the use of big data and statistical scoring processes can lead to grave consequences, exclusion and various disadvantages. The use of such information in the USA, such as automatic face recognition being permissible in court has lead to much higher risks of wrongful convictions for African-Americans. This was caused by the systems being trained with white faces, which picked up more detailed differences between test subjects.
Maas postulated, that we need an equal rights participation with the same chances for everyone. We cannot allow personal dignity to be denigrated by IT, we cannot become simply an object for an algorithm.
He also wants a transparency imperative for algorithms. For example the ranking and hiding of articles in Facebook, would need to be transparently explained.
This is already law for the "analogue" world, he now wants a counterpart law for the online world.
