cross-posted from: https://lemmy.world/post/4977520

(Scope for agency loss means that its comes at low cost to switch to an alternative and intensity of potential harm means the amount of harm that is inflicted on the individuum if the algorithm is bad (e.g. for lethal automatic weapons it would be very high))

An article, which explains the risk categorizes and actions to ensure their save usage: https://www.econstor.eu/bitstream/10419/233726/1/rego.12369.pdf

Based on the categorizes by Katharina Zweig, I would argue that algorithmic curation systems belong to category 0 (as long as there are different algorithms to chose from, which would be the case due to the federated-nature of services in the Fediverse). This would mean that they would not need to provide transparency.

Centrally managed algorithms on the other hand would belong to category 1 or 2 in my opinion and would require different control-mechanisms.

Therefore, in my opinion, the Fediverse could actually enable a healthy usage of social-media-algorithms.

(Scope for agency loss means that its comes at low cost to switch to an alternative and intensity of potential harm means the amount of harm that is inflicted on the individuum if the algorithm is bad (e.g. for lethal automatic weapons it would be very high))

An article, which explains the risk categorizes and actions to ensure their save usage: https://www.econstor.eu/bitstream/10419/233726/1/rego.12369.pdf

Based on the categorizes by Katharina Zweig, I would argue that algorithmic curation systems belong to category 0 (as long as there are different algorithms to chose from, which would be the case due to the federated-nature of services in the Fediverse). This would mean that they would not need to provide transparency.

Centrally managed algorithms on the other hand would belong to category 1 or 2 in my opinion and would require different control-mechanisms.

Therefore, in my opinion, the Fediverse could actually enable a healthy usage of social-media-algorithms.