This re-drafted report lays out a fundamental rights-based approach to AI ethics. The purpose of the guidelines is to ensure trustworthy AI: AI that is lawful; ethical; and robust. Trustworthy AI is tantamount, for the HLEG, because without that trust then the uptake of AI technology will be hindered. This would be a missed opportunity since AI is characterised as a means to help aid human flourishing and achieving the UN’s Sustainable Development Goals.
The foundation of trustworthy AI are the 4 right-based principles:
- Respect for human autonomy
- Prevention of Harm
The 7 key requirements, that build on the 4 principles, to realise trustworthy AI are:
- Human Agency and Oversight
- Technical Robustness and Safety
- Privacy and Data Governance
- Diversity, Non-Discrimination, and Fairness
- Societal and Environmental Wellbeing
While I think this is a good start for the EU in regulating the AI industry, I have yet to make up my mind on whether a rights-based approach to AI ethics would be best. I suspect that this was a case of realpolitik, since it is a lot easier for the EU to build regulation upon their pre-existing system of rights. After all, when it comes to these sorts of principles, they are explicitly laid out in the absence of regulation (thereby exposing the assumption that the purpose of ethics is to pre-empt regulation, with a ‘job well done’ once the law has caught up).
On the one hand, I appreciate the importance of trustworthiness in AI… however, when discussing the ethics of AI, it should not be for the sole purpose of establishing and maintaining trustworthiness. By discussing the ethics (read: morality) of AI, we do ourselves and society a disservice to limit ourselves in this way.
When we discuss the ethics of e.g. abortion or euthanasia, we do not do so with the aim of ensuring its uptake. Likewise, in discussing AI ethics we should be debating its merits, risks, effect on our lives and wellbeing so that we come to reasoned and principled positions on the matter… So that we can strive to be good and moral.
Also, for some criticism of the High-Level Expert Group from the inside, read here.
The overrepresentation of industry and lack of ethicists in the group is criticised along with the dilution of the original 'Red Lines'.