Experts in the fields of education, crisis management, and artificial intelligence gathered in Washington, D.C., on May 10 to discuss school safety and the potential of AI, in combination with a school response team, to identify and screen for warning signs and at-risk behaviors before a threat arises.
“What our schools are looking for and what they need is to spot behaviors and to intervene to prevent something, and to reach students who are troubled and coming off the rails to make a difference in his or her life as well,” said Jeff Jackson, president of the Georgia Independent School Association.
Harrowing events such as the Parkland, Fla., school shooting and a similar incident at Virginia Tech in 2007 have compelled some vendors to work on developing resources around school violence prevention.
The field is now looking to newer technologies like artificial intelligence to detect adverse behaviors before they escalate to a school crisis, said Suzanne Loughlin, chief administrative officer and general counsel for Novume Solutions.
Such tools are not yet available in easy-to-use formats for schools and districts, which means the task must be outsourced, but Loughlin is hopeful that demand for such solutions will eventually drive prices down and make them more affordable for school officials across the country to adopt and scale.
School violence prevention elements
Loughlin said AI solutions work by combing and continuously monitoring publicly available data on social networks to detect “natural language” from personal accounts. The software then flags any terms or phrases that suggest a potential school attack.
The solution works in tandem with a “formula” of preventative measures that schools and districts can implement to curb school violence.
As part of this, they propose methods such as awareness training to help members of the school community understand behaviors that are exhibited by perpetrators. They also suggest allowing students to anonymously report suspicions via text, email, or voice so that school officials can respond accordingly long before an attack occurs.
“Most schools have several of these elements in place,” she said of the formula, except for the use of “open-source intelligence monitoring” to track behaviors in real time, a near impossibility for time-strapped school officials and to the untrained eye, she noted.
‘Hidden’ in datasets
Mina Lux, founder and CEO of Meelo Logic, has conducted research on the use of social media to track at-risk student behavior and found that “a lot of these children who are planning [a school-based attack] want to be heard … they want to leave a mark.”
This means perpetrators are likely using social networking platforms where their status updates are maintained indefinitely and are available to a wide, unrestricted audience, making it easily accessed by AI tools.
She said potential school threats are often “hidden” in these large datasets.
“We’re looking for patterns,” said Hart Brown, chief operating officer and executive vice president for Firestorm Solutions, a crisis management firm.
“The complexity of what we see requires artificial intelligence, [a tool that] can deliver the right content information to administrators at the right time” and with more precision than humans, he said.
Emily Ann Brown covers education technology and STEM education issues for LRP Publications.
Copyright 2018© LRP Publications, Education Daily®