In Indonesia, extremists finding low-tech ways to get around high-tech blocks: Think tank

In Indonesia, extremists finding low-tech ways to get around high-tech blocks: Think tank

Extremists have mounted a series of attacks against Christians and other minorities in recent years
Extremists have mounted a series of attacks against Christians and other minorities in recent years AFP/JUNI KRISWANTO

JAKARTA: Indonesia’s cooperation with social media giants to remove extremist content is improving, but the extremists appear to be finding low-tech ways around blocks, according to a new report from the Institute for Policy Analysis of Conflict (IPAC), released on Friday (Jul 27).

While authorities were able to address a spike in online violent extremist exhortations during a May riot of terrorist suspects detained at the headquarters of the paramilitary police outside Jakarta, the incident also showed how quickly extremists can transfer material to other platforms and mirror sites.

Six people were killed during the riot, including five police investigators and one prisoner. A sixth police hostage was rescued and police regained control after 36 hours, but not before appeals went out over social media to Islamic State (IS) supporters across the country to come to the aid of the rioters against authorities, who were called the "oppressors".

Indonesia militants
A group of mobile brigade policemen patrol near an armoured car at the Mobile Police Brigade (Brimob) headquarters in Depok, south of Jakarta, Indonesia, May 9, 2018. (Photo: Reuters/Beawiharta)

"Extremist use of large, semi-public channels on apps like Telegram is declining, in Indonesia as elsewhere, but highly encrypted private communication over social media is still a problem,” said IPAC analyst Nava Nuraniyah, in a press release.  

She noted that while artificial intelligence and other tools are helping to identify and remove pro- IS messages, Indonesian IS supporters are creating hundreds of backup accounts to keep the propaganda going.

Authorities last year temporarily banned encrypted-messaging app Telegram over concerns it was being used to spread radical and terrorist propaganda.

Telegram then agreed to cooperate with Indonesia on a set of measures to clamp down on such material.

READ: Commentary: Don't win the battle against terrorism but lose the war on radicalisation

Indonesia's Minister of Communications and Information, Rudiantara, also met representatives from Twitter, Facebook and Google to discuss the issue so that the companies can all work together with the government and stem the spread of negative content.

The report said that new restrictions, coinciding with defeats of IS in the Middle East and the weakening of links between IS media channels and their supporters in Indonesia, led to a decline in the use of large, semi-public Telegram channels to disseminate propaganda.

But the use of highly encrypted, private small-group and two-person chats over Telegram continues.

As of mid-2018, government and social media companies have stepped up their efforts to detect and remove extremist content by using artificial intelligence (AI) and other tools to trawl the web. They also trained their respective AI machines to anticipate new tactics such as better encryption or other camouflage technology.

According to IPAC, while such innovation is commendable, most Indonesian IS supporters are not technologically sophisticated. 

READ: More than 500 terror plots foiled in Indonesia since 2012, says police chief

Instead of responding with high-tech countermeasures, they simply create hundreds of backup channels and accounts, move their groups and channels regularly, and store terabytes of propaganda material across various platforms and devices.

They are also exploring new encrypted messaging apps to prepare for the day when Telegram is no longer usable.

ARTIFICIAL INTELLIGENCE TOOLS

In January this year, Indonesia's Ministry of Communications and Information launched a new“Cyber Drone 9” system to enhance its cyber patrol capacity.

According to the IPAC report, this system - unlike the old system - uses AI tools to automatically detect violations online.

The ministry in June reportedly claimed that the crawler machine was able to detect more than 22,000 cases of radical content in the two weeks after the prison riot and blocked 4,000 of them. 

A team of 58 people operates 24 hours a day to review information collected by the crawler from websites and social media accounts.

There has been some criticism related to privacy infringement, to which the head of the ministry’s investigation and enforcement team, Teguh Arifiyadi, said that Cyber Drone 9 is not equipped with technology that can mine personal information.

The IPAC report explained that the Cyber Drone 9 system relies on human analysts to verify negative content that is detected automatically by the AI or reported by citizens or other government agencies.

FACEBOOK DUBBED "LEAST COOPERATIVE" IN DRIVE TO MITIGATE ONLINE EXTREMISM

The report said that in March, Minister Rudiantara released a performance report of eleven social media platforms, saying that Telegram had a 100 per cent compliance rate because it blocked 110 accounts out of the 110 recommended by the government.

Facebook, according to the report, was labelled “the least cooperative” because it failed to take down 34 per cent of all content flagged by the government. The disagreements relate mostly to different standards for assessing hate speech, especially in the context of political campaigns. Disagreement over violent extremist content is rare.

IPAC notes that IS online groups have been doing their own counterintelligence by turning to other apps such as Tam Tam messenger, which have been dubbed by some media as a “complete copy of Telegram” as of early 2018.

Despite significant progress in the government’s partnership with the tech companies, pro-IS online activists still manage to use encrypted private groups with impunity, the report said.

Governments may have to rely on a combination of high-tech algorithms and human infiltration to achieve maximum disruption of violent extremist groups online, the report said.

"CLEAR-CUT POLICIES SHOULD FOLLOW"

For this combination to be effective, IPAC recommended authorities get clear criteria and political consensus on what constitutes unacceptable content.

In the report, IPAC suggested that the Indonesian government lead and direct the effort to find this consensus, as pledges of commitment to moderation and the state ideology of Pancasila are not enough.

The report explained that if the government has a policy that labels hate speech or support for violent extremism as wrong, then a series of clear-cut policies should follow, and this includes banning clerics who violate those guidelines.

IPAC in the report also called for closer cooperation between government, tech companies and civil society, saying that one useful form of cooperation would be to compile a database of real examples to distinguish among legitimate political opinion, hate speech and criminal incitement that can be shared in discussions in all institutions across all parts of the country so that teachers are clear on what is acceptable and what is not.

It also said that security agencies, in particular, should conduct inter-agency training under the guidance of the Communication and Information Ministry along with consultants from the tech companies to ensure a common understanding of the guidelines and how to deal with violators.

Where encrypted communication is concerned, the report said the Indonesian government’s best bet would be to integrate online and offline intelligence in a more coordinated fashion.

Source: CNA/aa(hm)

Bookmark