ZB ZB
Live now
Start time
Playing for
End time
Listen live
Listen to NAME OF STATION
Up next
Listen live on
ZB

Facebook using real firearms footage to develop auto-blocking AI

Author
Newstalk ZB / Derek Cheng, NZ Herald,
Publish Date
Wed, 18 Sep 2019, 1:09PM
Facebook has announced a number of initiatives to improve its capability to stop the spread of March 15-type content. Photo / AP

Facebook using real firearms footage to develop auto-blocking AI

Author
Newstalk ZB / Derek Cheng, NZ Herald,
Publish Date
Wed, 18 Sep 2019, 1:09PM

Facebook has been using real-life first-person shooter video footage to develop artificial intelligence that can more effectively auto-block the type of video that was livestreamed on March 15.

It also plans to direct New Zealand users who search for white supremacy pages to anti-hate groups to help them de-radicalise, a strategy already in place for US users.

The use of the shooting footage, supplied by the US and UK Governments, is part of a Facebook announcement today about how the social media giant is building capacity to stop the spread of the kind of video that was livestreamed on March 15.

Facebook has been widely criticised over the way it failed to stop the spread of the March 15 footage, which was uploaded 1.5 million times over 24 hours.

Facebook auto-blocked 1.2 million videos, but it is not known how many people viewed the remaining 300,000 videos.

Since then, it has announced several initiatives, including signing up to the Christchurch Call and being party to a tech industry-led nine-point plan to target online terrorist and violent extremist content.

The tech-industry's AI shortcomings were also highlighted on March 15. A Facebook official reportedly told a congressional hearing in April that the livestreamed footage didn't have enough gore to be blocked automatically.

"The video of the attack in Christchurch did not prompt our automatic detection systems because we did not have enough content depicting first-person footage of violent events to effectively train our machine-learning technology," Facebook said in a statement.

"That's why we're working with government and law enforcement officials in the US and UK to obtain camera footage from their firearms training programs – providing a valuable source of data to train our systems.

"We aim to improve our detection of real-world, first-person footage of violent events and avoid incorrectly detecting other types of footage such as fictional content from movies or video games."

The Herald understands this work has already begun and shooter footage from the US and the UK has already been sent to Facebook.

Facebook also announced it is widening its strategy to direct users towards support groups when they search for white supremacy-related content.

In March last year, Facebook started an online tool whereby US users searching for such terms are directed to Life After Hate, an organisation founded by former extremists that provides education and support.

An example of how Facebook is redirecting users away from white supremacy-related content. Photo / Supplied
An example of how Facebook is redirecting users away from white supremacy-related content. Photo / Supplied

Today, Facebook announced it would introduce the same initiative in Australia, where the alleged Christchurch gunman was a citizen, and Indonesia.

Users will be directed to EXIT Australia and ruangngobrol.id respectively.

"These are local organisations focused on helping individuals leave the direction of violent extremism and terrorism," Facebook said.

Facebook wants to roll the programme out to New Zealand, though when that will happen and which local group it will partner with is still being determined.

This is seen as an important response to the Christchurch Call, which noted how the business models of online platforms can funnel users towards increasingly radical content.

One aspect of the Christchurch Call was for tech companies to monitor how this occurred and to use counter-narratives to re-direct users where necessary.

Facebook is also updating its definition of terrorist organisations; the previous one focused on violent acts, whereas the new one recognises the threat of violence and the intention to coerce and intimidate.

"Some of these changes predate the tragic terrorist attack in Christchurch, New Zealand, but that attack, and the global response to it in the form of the Christchurch Call to Action, has strongly influenced the recent updates to our policies and their enforcement," Facebook said.

"The attack saw the misuse of technology to spread radical expressions of hate, and highlighted where we needed to improve detection and enforcement against violent extremist content."

In May, Facebook announced more stringent terms of use for livestreaming, and has also banned content supporting white supremacy and white nationalism.

Prime Minister Jacinda Ardern welcomed Facebook's new initiatives.

"Facebook's actions highlight some of the early success of the Christchurch Call and show real change is happening.

"There is more work to do, but we have made a good start."

Ardern is expected to announce progress on the Christchurch Call next week during her visit to New York.

Take your Radio, Podcasts and Music with you