Discussions about ethical principles for the development and use of AI have been at the forefront of topical issues for some time, and there is no shortage of guidelines, principles and recommendations. However, the efficiency of their implementation remains questionable.
AlgorithmWatch, a Berlin-based NGO, analysed over 160 existing ethics in AI guidelines. It found that only ten of them included concrete proposals for enforcing the ethical behaviour, and that both private sector and public sector policies were “mostly voluntary commitments or general recommendations”. The analysis points out the vagueness of most ethics guidelines as well as the incompatibility of principles and practice. The NGO has also compiled a handy repository of ethics in AI guidelines from around the world.