Important points for #chatcontrol:

1) Is it all actually someone being abused?

Around 90% of reports made by PornHub in 2020 were not deemed child porn by the Canadian Police.

Around 90% of reports passed to Switzerland were not deemed to be criminally relevant.

Around 60% of reports passed to Ireland were not deemed to be "CSAM", although it's not clear whether this is an actual difference, or if the police are avoiding giving the benefit of the doubt.

The U.K. has reported they receive many reports which are not actually child porn.

Keep in mind the police have every incentive to play up the number of reports which contain it, or to use "clever" definitions to classify content which isn't it as it.

News publications reported there were many false positives as early as 2019.

2) Duplicates are prevalent. Facebook reported in 2021 the majority of images involved the same ten images slightly modified.

3) An internal study carried out by Facebook where they sampled x images revealed many things classified "exploitation" were posted for "shock" or as "memes". ???

4) In 2021, the Ninth Circuit Court of Appeals ruled the Fourth Amendment was violated when Google sent a report to NCMEC and NCMEC sent it onto law enforcement without a human (at either of them) reviewing the content.

That means that Google was *not even bothering to look at incidences* (or not all of them) and may still not be now in a number of cases.

* This incidence was automatically detected by Google's internal systems.