I am simply responding to the comment above that asks, who are the people who are requesting the information, and what is their agenda. That is also a very important thing to know for a vibrant public discourse.
I just can't help but think that pfizer did this on purpose - not making redacted versions of their data, knowing it would take the FDA forever to do that if it ever got FOIA'd.
This vaccine is a new type of medicine that was produced faster than vaccine before it. It also seems to work. But regardless, we all deserve to see every bit of information about its production, especially now that people's jobs depend on them getting it.
The CDC blew a lot of its credibility with the (arguably noble) early lie about masks not working.
Pharma companies probably don't have that much to lose.
Now it's the FDA's turn.
"We must hide legally-required-to-be-available information because some people may misuse it" is a path with unlimited abuse potential, best avoided.
There’s a good case to be made that releasing this is the best course. But let’s not pretend that public trust won’t be eroded because it takes an order of magnitude more work to disprove bulls### than to make it up (or twist it) in the first place. And it isn’t any particularly information I’m talking about, simply the vast quantity involved.
The vaccine becoming political is probably the worst thing to happen to humanity in the last 2-3 years. Worse probably than COVID itself.
EDIT: just thought of one possible solution or at least mitigation to the “Bull### asymmetry principle” when combined with massive data dumps: Limit the amount of material that can be requested. Any specific dates or whatever could be requested within the stack. Or if a representative sample is desired, the entire dump is sectioned in pieces and each section is securely hashed. The one doing the request can then request, say, up to 100-1000 sections and each such section can then have its hash compared. That ensures the random selection is random to the satisfaction of the requesting party. And it limits the amount of bulls### that needs to be countered to some reasonable level. Otherwise, any attacker will be able to overwhelm the public with endless claims from a massive stash of requested documents.
This doesn’t limit the amount of work for scrubbing PII, though, since you’d have to scrub it before hashing or the hashing wouldn’t be recreatable. But at least you’d limit the amount of unique bulls### that’d need to be countered.
Really, I see FOIA requests as failures of government; as in they represent the government being poked to be transparent when transparency should be the default.