Social Sentinel gives Flagler Schools a false sense of security
First I want to say thank you to School Board attorney Kristy Gavin and Flagler Schools Superintendent James Tager for their prompt reply to my public records requests and making the process painless. The information below is based upon the alerts received by Flagler Schools via the Social Sentinel platform for May, June and July.
As you may recall, Flagler Schools approved an over-$18,000-per-year contract with Vermont-based Social Sentinel on April 17 this year. The school administration and School Board claim the service is to help protect the schools and individuals by the monitoring of social media posts using a technique called geofencing. The funds used are from the half-penny sales tax referendum for technology in education.
There were 25 alerts generated (one appears to be deleted by the user), and all far from harmful, to be honest. Sheriff Rick Staly appears twice.
All of the alerts were from Twitter — only Twitter. This is probably the most significant and telling piece, in my opinion, to the Social Sentinel platform and its value, or lack of value.
Twitter ranks third in terms of the number of users in the United States (approximately 207.36 million Facebook users and 96.3 million Instagram users, compared to the 64.9 million Twitter users, according to Statista.com). Facebook and Instagram typically rank much higher in active users as well. If Social Sentinel is only monitoring Twitter, its value is limited.
So why is Social Sentinel not discovering and generating alerts from other sources? Because Facebook and Instagram (a Facebook-owned company) do not allow the use of their feeds for conducting surveillance.
Social Sentinel claims not to conduct surveillance, but their platform in my opinion is the very definition of “surveillance,” which is defined by Merriam-Webster as the “close watch over someone or something.”
Unless a post is made directly to the Flagler Schools Facebook page, Social Sentinel can not see the post, whether it is public or not.
Who is being monitored?
The other question that has been on inquiring minds is whether the service was monitoring the entire county or just the school grounds. Flagler Schools provided the specific geolocations (longitude and latitude) of the posts in May and June, and it appeared that the results were not contained to just school grounds. In the July batch, only the number of miles from a school was provided, and most were within 5-6 miles of several schools listed on each alert. One tweet from @therob424 that is questionable seems to be listed as being 567 miles from the schools listed on the alert.
Why the status quo is dangerous
So is the service worth the cost? No. In fact, it’s actually more dangerous to put any reliance on the service because it creates the false impression the schools have a “leg up” on being alerted to threats, yet they really only have access to a small fraction of activity online.
This was obvious in the “Live” interviews on FlaglerLive.com with Trevor Tucker and Janet McDonald. Tucker said, “I feel that Social Sentinel is a valuable asset to the school district, in that it can screen millions of different social sites and posts to make sure students are not trying to create violence to others or themselves, or spread messages of hate.”
McDonald said, “The district’s contract with Social Sentinel is a very economical service, scanning public information on social media for imminent threat and personal harm indicators. Phrases or ‘hits’ have been researched and cleared. So we know that the system is working for the items we’ve identified as primary safety topics.”
Obviously, neither has followed up with the schools since voting in favor of the service and are under the impression it is delivering more than it actually is. My opinion is that more investment and focus by Flagler Schools on engagement and conversing with the community online rather than trying to conduct surveillance would be a far more effective approach.
Brad West lives in Palm Coast.