Here is an idea for better oversight of big tech that does not involve more regulation: a deacentralized moderation system that replicates the us court system.
If something is flagged, then a jury made up of randomly selected users decide if the comment is inciting violence / etc.
If 1 user says it does not , then there is no ban/deletion etc. You can iterate many times off that into a system that is scalable, cost effective, descentralized, transparent, and just.