Amazon late Friday disabled a website used by a propaganda arm of the Islamic State that celebrated the suicide bombing that killed at least 170 people in Kabul on Thursday after The Washington Post reported the extremists relied on the company’s technology to promote extremism.

Nida-e-Haqq, an Islamic State media group that distributes Islamist content in the Urdu language, had been using the company’s dominant cloud-computing division, Amazon Web Services, to host its content, despite company policies against working with terror groups.

Some of that content included messages about the Islamic State-Khorasan offshoot that claimed responsibility for the lethal attack, said Rita Katz, executive director of SITE Intelligence Group, which monitors online extremism and discovered the link with Amazon Web Services. Urdu is widely spoken in neighboring Pakistan and occasionally in Afghanistan itself.

The Nida-e-Haqq app on Thursday carried what it claimed was an image of the bomber wrapped in a suicide vest ahead of a blast whose victims included 13 U.S. service members, further marring the American pullout from the nation after nearly 20 years of war.

” (F) ollowing an investigation, we have disabled a website that was linked to this app as it was in violation of the AWS Acceptable Use Policy,” Amazon spokesman Casey McGee said in an emailed statement send late Friday after The Post reported on SITE’s findings.

That policy bars customers from, among other practices, using the cloud-computing service “to threaten, incite, promote, or actively encourage violence, terrorism, or other serious harm.”

The source code for the app, Katz said, shows it draws words and images from a website for the Islamic State propaganda arm. That website, whose content is password protected and could not be directly reviewed by The Washington Post, has been hosted by Amazon Web Services since April, according to online domain records.

“It’s just mind-blowing that even after all these years, ISIS could still find a way to exploit a hosting company like Amazon,” Katz said. “Of course, we should presume that ISIS will always be searching for ways to bypass security protocols, but this app isn’t even trying to stay low-key. It is blatantly filled with official ISIS claims, media and logos of ISIS’ media arms, clear as day. This app was clearly created to keep ISIS’ message and content alive and distributed online. . . . It is clear that the stakes of keeping such content offline is no less major than in past years.”

Amazon, like most American technology companies, has policies prohibiting Islamist terror groups from using its services. (Amazon founder and executive chairman Jeff Bezos owns The Washington Post.)

Amazon took a stand in January against Parler, a social media site popular with supporters of former president Donald Trump, severing ties over the site’s failure to adequately monitor hate speech and calls for violence related to the Jan. 6 siege of the U.S. Capitol. The move knocked Parler offline for weeks and sharpened political debate over the power of tech companies to determine what appears online.

But enforcement of the policy appears to have failed in the case of Nida-e-Haqq.

One reason may be that Amazon hasn’t often proactively policed the content of customers – many of which run widely used commercial websites such as Airbnb, Yelp and Netflix. Rather, its Trust & Safety team, which has fewer than 100 workers, acts only on complaints received.

Despite the size of that team, Amazon is the dominant provider of cloud infrastructure services, which let customers rent data storage and processing capabilities over the Web instead of running their own data centers. AWS, which competes against Microsoft and Google, held 41% of the global market in 2020, according to the market research firm Gartner.