Highlights are as follows:

• they will be “fast-tracking” their decision to force apps and sites (“tech firms”, of course, because all apps and sites are “tech firms”) bringing the decision forward
• their proposal will be that everyone be forced to use fuzzy-image-matching, which they dress up as “perceptual-” or “hash-matching”, but it’s an old tech wrapped up in legal agreements with inadequate oversight
• they frame this as a means to combat non-consensual deep-fakes, even though those are not currently a large part of the extant “hash databases” in use by the likes of Meta, which are geared at CSAM and Terrorism
• management of such databases, and likewise logging of the queries where those databases are offered as an “outsource service” — clearly this is where the money will be made — both lead to major security issues; this is not mentioned
• they will decide in May and expect other legalities to be sorted by “this summer”