Local experts say monitoring online posts for potential threats a difficult task

BOSTON (WHDH) - In the wake of deadly mass shootings in El Paso, Texas and Dayton, Ohio, President Donald Trump is calling for greater policing of online posts, but local experts say that might not be as easy as it sounds.

“We must shine light on the dark recesses of the internet and stop mass murders before they start,” Trump declared during a news conference on Monday.

The suspected El Paso shooter is believed to have posted a hate-filled manifesto on an internet message board prior to the shooting that killed 22 people.

RELATED: Authorities scour leads in mass shootings that left 31 dead

“The perils of the internet and social media cannot be ignored and they will not be ignored,” Trump said.

He directed the Department of Justice to work with police departments and social media companies in order to develop tools to detect potential mass shooters before they strike.

Computer programs already monitor public posts for key words, phrases and hashtags but Joan Donovan of the Harvard University’s Shorenstein Center said one challenge is analyzing and predicting what’s a real threat.

“You could just as easily say something like you know, ‘I want to kill that guy’ and not mean it,” she explained.

Internet security expert Joseph Steinberg added that the government would have difficulty monitoring private posts or personal chats.

“Scanning those kinds of posts today would require in many cases a warrant and that’s not a warrant the government will get en mass,” he explained. “They’re not going to be able to say, ‘Let us be able to scan everybody’s post because we suspect everybody may be a shooter.’ That’s just not going to happen.”

Social media companies allow users to report posts to the network.

Experts urge users to call police if they spot something online that they think needs immediate attention.

(Copyright (c) 2019 Sunbeam Television. All Rights Reserved. This material may not be published, broadcast, rewritten, or redistributed.)