Facebook user Jonathan Spyer, a respected journalist and academic, had his profile banned on the social network because he mentioned the threat radical Islam poses to Europe. The company reinstated his account later but did not give an explanation, reports the Daily Caller.
Why did Spyer get banned?
Last month in the original post provided to The Daily Caller News Foundation, Spyer wrote, “The Islamist insurgency in Europe continues. Here are some preliminary thoughts from yours truly regarding the inability of mainstream western elites to process what is occurring.”
The post included a link to the post on his blog.
Several hours later, Spyer received a notification from Facebook that his account had been “disabled” for stating that the recent terror attacks in Germany and France essentially constitute a “low-level Islamist insurgency.”
Mentions of “Islamist” and “nsurgency” not acceptable
Spyer, an analyst at The Middle East and senior research fellow at the Global Research in International Affairs Center, wrote a brief post noting that keywords such as “Islamist” and “insurgency” apparently led to the triggering of Facebook’s post review mechanisms, eventually leading to a ban on his account. Spyer noted the ban in a post last week on his blog, the report says.
Much to everyone’s surprise, Facebook only recently reinstated Spyer’s account. No reasons for the reinstatement were given, but it could be because of the attention grabbed by Spyer’s post against the ban.
“[T]he process by which the profile was destroyed is interesting and may be informative regarding the practices of Facebook with regard to the issue of freedom of expression on the site,” Spyer told TheDCNF.
Facebook open-sources AI-building research
In other Facebook-related news, the company has decided to open-source its AI bot-building research — part of Facebook AI Research (FAIR) Lab’s mission. The company is planning to provide better access to researchers and engineers to help them build better bots. The code library called “fastText” is now available on Github, and Facebook says it just requires a compiler with “good C++11 support.”
The usefulness of fastText is in the efficiency and speed at which one can train a bot. One can train models “On more than 1 billion words in less than 10 minutes using a standalone multicore CPU,” claims FAIR. Training time can vary from days to seconds in fastText, compared to deep learning models.
Facebook integrated chat bots in its Messenger app earlier this year, and this makes it easy to see why this move makes sense for it.