Published on the 26/09/2019 | Written by Heather Wright
Just how many moles are being whacked with online hate prevention tools?…
Digital ‘fingerprints’ related to the Christchurch terrorist attacks accounted for over 1,000 of the unique hashes collected by the Global Internet Forum to Counter Terrorism (GIFCT), a new report has revealed.
The GIFCT, founded in 2017 by Facebook, Microsoft, Twitter and YouTube as a co-ordinated group to help shut down extremist content across platforms, this week announced it is forming an independent body, with its own dedicated staff as it continues to try to thwart efforts by extremists to ‘abuse digital platforms’.
Part of that effort lies with the shared industry hash database. The database of photos, videos and URLs is shared among GIFCT’s members, which now include Dropbox and Pinterest, and the Hash Sharing Consortium – a group that predates GIFCT – in order to more quickly shut down the spread of extremist content. The hashes can be used to identify and remove matching content that violates members respective policies or, in some cases, to block terrorist content before it is posted.
The rapidly growing hash database also provides a means to classify and measure the extremist content that is being identified. For the GIFCT, the metrics are central to demonstrate to inquisitive world leaders (and more importantly, the regulators and policy makers under them) that something is being done.
The new, independent, status and governmental oversight doesn’t however confer any new teeth.
It’s a sad question, but what’s the most popular offensive genre?
‘Glorification of terrorist acts after the fact’ made up 85.5 percent of all content in the shared database of hashes, with ‘imminent credible threats’ – the public posting of a specific credible threat of violence yet to happen – accounted for 0.4 percent and was the smallest category in the Transparency Report.
‘Radicalisation, recruitment and instruction’ – deemed to be any material seeking to recruit followers, give guidance or instruct them operationally – accounted for 9.1 percent of content.
‘Graphic violence against defenceless people’, such as murder, execution, rape and torture, accounted for 4.8 percent (that’s some 9,600 unique instances if we assume just 200,000 hashes have been collected).
Hashes collected by GIFCT classification:
• Imminent Credible Threat: 0.4%
• Graphic Violence Against Defenseless People: 4.8%
• Glorification of Terrorist Acts: 85.5%
• Radicalization, Recruitment, Instruction: 9.1%
• New Zealand Perpetrator Content: 0.6%
As has been widely reported, the hash database was already in place, but failed to stop the spread the live footage of the Christchurch attacks. Facebook removed more than 1.5 million videos within 24 hours of the attack, but still some 300,000 viewers saw the footage from versions remaining online for much longer. GIFCT notes that the ‘extraordinary virality of the attacker’s video online illustrated the need to do even more’.
That ‘even more’ appears to be the development of the independent body, which will have an independent advisory committee, chaired by a non-governmental representative, and include members from ‘civil society’ – that’s groups outside of government and the business sector – government and inter-government agencies. New Zealand has signed on to the advisory committee, along with the US, UK, France, Canada and Japan but Australia is currently missing in action.
“The new independent GIFCT will integrate its existing work to develop technology, cultivate strong corporate policies and sponsor research with efforts to fulfil commitments in the nine-point action plan released after the Christchurch Call,” the GIFCT says.
The new, independent, status and governmental oversight doesn’t however confer any new teeth. It still relies on GIFCT members voluntarily ‘enforcing’ their own policies and standards.
Amazon, LinkedIn and WhatsApp are also joining the GIFCT, it announced this week.