In the age of artificial intelligence (AI), understanding the role of politics concerning AI is crucial. Researchers Eva Erman and Markus Furendal emphasize the distinction between "governance of AI" and "governance by AI" as separate phenomena requiring careful consideration.

The term "governance of AI" refers to the development of frameworks, policies, and regulations that guide and control the use of AI systems. It involves human decision-making and oversight in shaping the rules and principles governing AI technologies. Examples include the EU's proposed "AI Act" and various soft law approaches such as recommendations, ethical guidelines, and codes of conduct developed by different stakeholders.

On the other hand, "governance by AI" describes the implementation of AI systems within existing governance mechanisms. This involves using AI for decision-making, data processing, and automation in public and private institutions. It ranges from recommendation algorithms and chatbots to automated decision-making by public authorities. While humans are still involved in decision-making processes supported by AI, further delegating decision-making to machines is possible.

The authors argue that it is essential to distinguish between these two notions of AI governance. They suggest that the legitimacy of AI governance is not solely determined by its outcomes but also by the procedures and democratic participation involved in its decision-making processes.
The distinction becomes particularly important when considering the normative aspects of AI governance, such as its impact on democracy.

The governance of AI can be politically legitimate when there is a democratic chain of legitimacy, where decisions are made by authorized entities accountable to the citizens they represent. For example, the EU's AI Act derives its legitimacy from the democratic procedures within the EU. In contrast, governance by AI initiated by private companies lacks this democratic legitimacy, even if their initiatives aim to promote responsible AI.

The authors also question the legitimacy of governance by AI systems in the stronger sense, where AI would become an authorized entity making decisions on behalf of humans. They argue that democracy not only aims for good decision-making but also emphasizes self-determination and the equal participation of those affected by the rules. Handing over ultimate authority to AI systems would undermine political legitimacy.

In conclusion, Erman and Furendal emphasize the need for a nuanced understanding of "AI governance" and its implications. They highlight the normative and practical questions raised by both governance of AI and governance by AI, stressing the importance of studying and developing theories that align with the ideals of social and political interactions in the context of AI technology.

Thanks for reading. You can support and reward my writing via: 

Pay Pal – [email protected]

Algo - NCG6LBALQHENQUSR77KOR6SS42FGK54BZ5L2HFDSBGQVLGYIOVWYDXFDI4

ADA – addr1q9vfs6nqz4xmtnpljwhv4tukyskd2g7enxd87rpugkwwvfun5pnla5d5tes2mvurrc77e7837yd0scrfk063qlha8wgs8d4ynz

Bitcoin 3HbxyDXE9MhNQ8RqsirqgYvFupQzh5Xby2

ETH - 0x8982cdb97bd23f092f78a16a4fc93c5c4607a285

Seeds – vladlausevic

Skycoin – ZxjhWMJRbTNCRQzy5MekZzH4fhdWFCqBP8

Tezos — tz1QrRzkTAKuPKF8dmGW6c1ScEHBUGvoiJBM