Many persons have been sending me materials relating to the death last week of 23-year-old Korryn Gaines during a violent police confrontation (in the process of serving a warrant) at her Baltimore area home. Of particular note in these messages has been Facebook’s decision to temporarily suspend her Facebook account about seven hours into the ongoing standoff when police asked Facebook to do so (her Instagram account was temporarily suspended as well).
Gaines had been recording videos of the confrontation and posting them as the standoff continued. Far more troubling were her followers, many of whom — in response to those videos — were apparently urging her not to comply with police and even suggesting aggressive actions against them.
Sometime after the accounts were suspended, police shot and killed Gaines, who was herself reportedly threatening police with a shotgun. Her 5-year-old child was also shot but is reportedly recovering.
The main reason I haven’t commented on this case publicly to date is that, frankly, I’ve been thinking about it and didn’t come to any immediate conclusions.
One way I try to analyze complicated Internet-related issues is to see if I can think of parallels in the “non-Internet” world that might shed some light on the matter.
Such parallels do exist in this case, and suggest that the most problematic aspect of the technology-related portion of this tragedy wasn’t the videos being posted per se, but rather the feedback Gaines was receiving from her followers in real time.
If we think about this situation in a non-Internet context — an angry confrontation, a suicidal person, or other similar scenarios — law enforcement would normally attempt to clear boisterous onlookers (“Go ahead, jump!” — “Shoot the pigs!”) from the scene, so that negotiations (in the case of Gaines, we’re talking more than seven hours) could proceed with some semblance of calm and without third parties attempting to escalate the situation for their own sordid jollies.
By these analogies, frustrated police in requesting the account suspensions were doing the social media equivalent of getting the yelling crowd away from the negotiation scene (which of course also has the effect of getting potential witnesses away from the scene, we must also note).
In this particular instance I feel that — overall — the police and Facebook/Instagram’s social media account actions perhaps were on balance justified, but that’s not the end of the story by any means.
We really need to often conceptually separate the videos themselves (being broadcast live over social media, or being posted in real time), from the live responses and comments that viewers of those videos are making back to the person in the confrontation itself, though this area is also very complicated.
For example, we’ve already seen cases of persons streaming live Facebook video to broadcast a suicide, and in another instance a rape. In such circumstances, it can certainly be argued that the videos alone are egregious enough to warrant blocking.
But it’s the instant feedback aspect of comments and chat dialogues — typically associated with live or posted videos — that seem the most problematic in ongoing confrontations, in the same manner as the crowd screaming for blood outside a physical building.
This all suggests to me that society, law enforcement, and the social media firms themselves would benefit in the long run from a more finely-grained set of tools to deal with these these kinds of events.
We can start with the given that cutting off a person’s social media accounts at the request of law enforcement should always be a last resort only to be used when absolutely required — not a first-order default decision.
But when the decision is made to take actions in this regard, there may be many instances where simply cutting off the feedback to the user rather than shutting down the videos and entire account may be more appropriate — the equivalent of getting the screaming crowd pushed back for a time so that negotiations can proceed with less chaos.
Would the user become angry or upset when they realized that the real-time feedback had ceased? Perhaps, but probably less angry or upset than they’d be if the entire account suddenly went dark.
We’re on the cusp of a vast explosion in the numbers of these kinds of situations in which social media will play important, even crucial roles. Today the policies and tools for dealing with these events appropriately are either too primitive and coarse, or simply don’t really exist at all.
We have a lot of work to do.
–Lauren–
I have consulted to Google, but I am not currently doing so — my opinions expressed here are mine alone.
“We can start with the given that cutting off a person’s social media accounts at the request of law enforcement should always be a last resort only to be used when absolutely required — not a first-order default decision.”
This is where I think some balance needs to be found. I’m not too thrilled at the idea that the police can contact communications providers, who are otherwise uninvolved, and request that someone’s service be shut down. At the same time, I can understand the request and why law enforcement made it. If we look back over the decades, one method that has been repeatedly used to end standoff situations is to cut utilities (electricity, water) to the target. But seige tactics usually don’t come into play until a standoff stretches from hours into days, and there are various examples of how it didn’t work out too well. Ruby Ridge, Waco, and the Bundy dorks all come to mind.
Yes, there were apparently people who were encouraging Gaines to commit illegal activities. But there may also have been people trying to contact her to encourage her to stand down and surrender. Such messages couldn’t get through because the police had frozen her social media accounts. It seems to me that the more logical move would have been asking Facebook to temporarily suppress messages from those people who were inciting trouble. To use an analogy, if someone is threatening suicide by jumping off a building, you don’t cut that person off from the outside world; instead, you probably want to address the instigators who are yelling “Jump! Do it!”
In recent years, law enforcement has taken to escalating situations without necessity and far too soon. This, I believe, is the real problem.
Seven hours in a busy populated area strikes me as a pretty long negotiation with someone who has a small child (effectively being used as a hostage) and a shotgun. I’ve seen little evidence of good outcomes with days-long negotiations. In the examples you cite it seems as if the extended periods were mainly used by the perpetrators to prepare violent ends. I don’t see any way to reasonably allow only the “good” people to make comments and somehow filter out the “bad” ones. Normally what you want to do is get the outside chaos away so that you have a clean 1:1 negotiating environment. Can you imagine the liability exposure to a social media firm that attempted to cherry-pick comments it would let through in a situation like that?
“I don’t see any way to reasonably allow only the “good” people to make comments and somehow filter out the “bad” ones.”
I don’t really disagree. One curiosity of mine is why did law enforcement approach Facebook specifically instead of going to her cellphone carrier and asking them to suspend all data? Is it likely that the cellphone carrier would have demanded a warrant or court order, or some other legal authorization beyond a city police department just making a request?
“Can you imagine the liability exposure to a social media firm that attempted to cherry-pick comments it would let through in a situation like that?”
I wasn’t suggesting that any social media provider should be the entity making that call, and I think we probably agree here for the most part in that there are boundaries that need to be set. I’m just uncomfortable with the idea of suspending someone’s communications because third parties might be suggesting illegal activity or the furtherance of a criminal act.
Much more complicated to deal with the carriers, and less likely to be definitive. She could have been on Wi-Fi, or switched to Wi-Fi if the carrier data was dropped. And you want to keep the voice call capability open, which can be complicated especially on LTE phones if you want to block the data per se. Point being it’s much more hit or miss when the event clock is running. I agree that such actions on *suspicion* are much more problematic. However, in this case, since she had responded to a legal warrant by barricading herself and her child for more than seven hours, they had her with almost 100% certainty on a variety of charges related to those acts alone.