The removal of red lines

[Shishirgano9]

As a member of the expert group, I am disappointed with the outcome presented. The guidelines are lukewarm, short-sighted and deliberately vague. They ignore long-term risks, gloss over major issues (“explainability”) through rhetoric, violate basic principles of rationality and claim to know things that no one really knows.

Together with the excellent Berlin machine learning el salvador mobile database expert Urs Bergmann (Zalando), my task was to develop, over several months of discussions, the so-called “red lines”, non-negotiable ethical principles that determined what should not be done with AI in Europe. The use of lethal autonomous weapons systems was clearly on our list, as was the state’s assessment of citizens (social score) supported by AI and, in principle, the use of AI that people can no longer understand or control.

I only realised that this was not really wanted when our friendly Finnish HLEG chairman Pekka Ala-Pietilä (a former Nokia employee) asked me in a soft voice if we could remove the phrase “non-negotiable” from the document. In the next step, many industry representatives and group members interested in a “positive vision” insisted vehemently that the “red lines” be removed from the text altogether – even though it is precisely these red lines that fall within our mandate. The published document no longer mentions “red lines”; three have been completely removed and the rest diluted. Instead, only “critical concerns” are mentioned.