ALABAMA - After the deadly shooting Saturday at a Pittsburgh synagogue, federal law enforcement officials discovered the shooter made anti-Semitic statements during the event and targeted Jews on social media. Which brings up the question: Is hate speech protected by the First Amendment?
Reports say Robert Bowers frequently used the fringe social media site Gab and posted anti-semitic messages there, including one right before the attack. Gab has a little more than 200,000 users and is known for having few restrictions for content that is posted. Monday the Gab site was offline. Many companies are distancing themselves from the site including its web hosting company Go Daddy which gave Gab 24 hours to find a new host provider.
In place of the Gab site, CEO Andrew Torba has left a letter. In it, he says Gab isn't going anywhere and he will continue to fight for free speech. But what is considered free speech? And how is this constantly being redefined online?
Go Daddy has pulled the plug on Gab as several other companies are distancing themselves from the site. While the CEO of Gab says his site protects free speech, are these companies violating any laws by separating themselves from it?
"The First Amendment only says that the government cannot intervene and stop our freedom of speech," constitutional law attorney Leroy Maxwell said.
In this situation, he says the mighty dollar reigns supreme.
"The government didn't step in. There's no violation here of overstepping the First Amendment. This is the private company, Go Daddy, who happened to be the server for Gab, who decided to jump in and say listen you're bad for business."
Billions of people post on social media every day and sites like Reddit, 4chan, and Gab provide a place where people like the Pittsburgh shooter can post messages filled with hate. Are these posts protected under the First Amendment?
"Its protected. Hate speech is protected unless it crosses over the lines of speech that becomes one that can incite violence," Maxwell said.
Maxwell says making a violent threat is where the government can step in, but proving if a post is inciting violence is difficult for many social media platforms including Facebook, Twitter, and Instagram.
"They're dealing with lawsuits when it comes to cyber-bullying and other folks committing suicide because of people's words, and so the line is constantly evolving, and constantly changing where hate speech turns from protected to inciting violence," he said.
Because of that, these sites are having to redefine the line of what kind of speech is protected and what is not.
How is social media monitored?
Everyday people log in to social media and post their thoughts, opinions, and sometimes even violent sentiments.
"So, there's a number of cooperations as well an initiative between the public and private sector to better monitor social media, especially social media, in order to identify these types of activities and sort of flag them as potentially malicious," PeopleSec Chief Hacker and CEO Joshua Crumbaugh said.
Many companies are using algorithms to detect danger.
"We put our trust in these big data machine learning algorithms and it's misplaced," Dr. Candice Lanius said.
Dr. Lanius is an assistant professor at UAH specializing in social data. She says algorithms can tell the difference between certain words like cat and dog, but in other areas, they can fall short.
"If its the difference between an edgy joke or something that is the valence and the intention behind it is hateful, the machine is not going to be very good at identifying that."
"None of these tools are incredibly accurate and they produce a lot of alerts for people that may never be violent or never committed a crime," Crumbaugh said.
Which poses a question:
"How do we differentiate between these people that just like to, I don't know, go out on the internet and say these horrible things just to see what kind of reaction they're going to get, and then the people who are saying these things because they truly believe them and now they might act on it," Crumbaugh said.
And coming up with an answer won't be easy.