Today’s NY Post makes an alarming claim, according to FBI whistleblowers. They state that Facebook gave the FBI posts from conservatives unhappy about 2020 elections without probable cause. This triggered numerous investigations.
The Post article has some great details. I love the complaint by the agents that they received too much data on people not actually threatening them, which was a waste of investigative resources. You can’t get more intrusive, but not productive make-work to encourage FBI agents to be civil libertarians and expose surveillance programs.
The story has one problem, however. This conduct would go against the law and would not be tolerated by Facebook or the FBI. Electronic service providers are generally prohibited from providing customer information to government agencies under federal law.
Facebook also issued a denial. A very careful denial. This statement states that it is simply not true and has zero support.
Such a compound denial can often mean that only a portion or minor variation of the statement is true. The denial of Facebook screening is not valid if it is looking for “anti-government language” or questions concerning the legitimacy of previous elections.
In an attempt to reconcile the story with the denial, the Post suggests that the FBI may have recruited a Facebook employee (CHS) as a confidential source of human resources. That is doubtful. However, being a CHS means you cannot do any things with the employer’s information that your employer isn’t able to. It is unlikely that the FBI will be free to bypass a certain limit in its investigative powers by using CHS.
However, federal law provides that electronic service providers can provide any information they wish to law enforcement. To do so, they need to believe “in good faith … that an emergency involving danger of death or serious physical injury to any person requires disclosure without delay of communications relating to the emergency.” 18 USC 2702 (c).
Facebook and Silicon Valley firms could have designed an AI engine for searching strings of words its legal team has cleared — in good faith — to indicate an emergency that poses a risk of serious injury or death. The injuries sustained in the January 6th riot can be used as evidence to forecast such dangers from “rigged” election talk and antigovernment.
You could extract these passages from social media sites, together with your identifying information and send them to Justice under the exception for “danger of injury or death”. Justice could use these to request all less inflaming posts from the same individuals and send the results to the local FBI office for an investigation.
Important note: It is impossible for me to know if any of these are happening. Just trying to discover a legal route that could make the Post stories more plausible. My program above will better reflect the facts, which include the Facebook denials as well as the absurdity that Justice and FBI have flouted the law.
Just because something is legal, doesn’t necessarily mean it’s good. As all content moderation students know, any mass effort to identify “bad” speech across a large social media platform will make many mistakes.
It is not surprising that mass Silicon Valley criminal referrals would favor conservatives as well, just like content moderation. This bias could be exacerbated if Justice uses an existing grand jury that was established on January 6, to issue subpoenas.
Assuming I am correct, one might wonder how such an effort was created, what aggressively conservative complaints turned into emergencies threats of life and limb, and who oversees the process to stop overbroad seizures legitimate speech. It is also possible to inquire whether Black Lives Matter or any other movements whose extreme supporters have occasionally fallen into violence.