Requirements from EU Digital Services Act

[esapulkkinen]

Notice that EU Digital Services Act ( https://eur-lex.europa.eu/eli/reg/2022/2065/oj ) requires adherence to EU fundamental rights when making moderation decisions.

Couple of important points:

• Terms of service cannot violate fundamental rights, and cannot require violations from users. E.g. asking for "all rights reserved" would violate this.
• It is probably useful to distinguish when EU DSA is applicable, if this is feasible. However when it's not, probably more general human rights considerations should apply.
• Fundamental rights assessments should consider fundamental rights of all actual persons. Robots or AI bots however do not have same rights.
• EU countries are considered similar by EU rules, but national rules sometimes provide additional rules.
• Cross-border situations are more complicated than situations where rules of only one country apply.

[ghost]

article 15 Transparency reporting obligations for providers of intermediary services any use made of automated means for the purpose of content moderation, including a qualitative description, a specification of the precise purposes, indicators of the accuracy and the possible rate of error of the automated means used in fulfilling those purposes, and any safeguards applied.

[aloiscochard]

@ernkurm thanks for extracting the relevant article.

Could you please transfigure this in mundane terms?

This would be helpful to all the non-lawyers part of the open source community.

@esapulkkinen It seems you really wanted to share that link here, but do you have any actionable items for it? I don't know something you discussed in your own little circle maybe?

[ghost]

hi thanks for reaching out and I have not transfigure this in mundane terms , so not really helpful for open source community. I don't have any and have not my own little circle.

[esapulkkinen]

@aloiscochard It's slightly difficult to construct actionable items, since there is significant uncertainty as to what is technically feasible. However, my personal opinion is that following should be considered:

When moderating content, different countries have slightly different requirements with respect to what kind of content is considered (un)acceptable. So when the algorithm decides to flag content for moderation decisions, that locale-specific variation in rules could be used as one criterion when information about the location/country of users is available. This might trigger additional filtering rules when transferring data between countries, and allows applying local rules within particular locale, e.g. in Europe there are rules concerning distribution of personally identifiable information (e.g. due to GDPR requirements). Of course the full set of applicable rules is somewhat complex, and the rules are definitely different in different countries. I suspect automated flagging of content can improve possibility that such distinctions can be automatically implemented. It's not completely clear to me if this idea is feasible.