For technologies with a high degree of automation, an extended application of the precautionary principle is necessary in order to proactively avoid stress or damage or to reduce them as far as possible.
 
The examples listed show various ethical problems. They cannot be resolved only with high-level ethical principles. The principles must be reflected in specific instructions.

The algorithmic implementation of a fairness principle must be preceded by a discussion on the theory of justice.
 
On the long-term view, partnerships between universities and companies are a viable option. Joint projects with research institutes are also possible. In this way, companies can build up the ethical expertise they need to design artificially intelligent systems in a generally compatible manner.
 
You can read the full analysis here as a pdf.