An update on our commitment to fight terror content online
A little over a month ago, we told you about the four new steps we’re taking to combat terrorist content on YouTube: better detection and faster removal driven by machine learning, more experts to alert us to content that needs review, tougher standards for videos that are controversial but do not violate our policies, and more work in the counter-terrorism space.
We wanted to give you an update on these commitments:
Better detection and faster removal driven by machine learning: We’ve always used a mix of technology and human review to address the ever-changing challenges around controversial content on YouTube. We recently began developing and implementing cutting-edge machine learning technology designed to help us identify and remove violent extremism and terrorism-related content in a scalable way. We have started rolling out these tools and we are already seeing some positive progress:
More experts: Of course, our systems are only as good as the the data they’re based on. Over the past weeks, we have begun working with more than 15 additional expert NGOs and institutions through our Trusted Flagger program, including the Anti-Defamation League, the No Hate Speech Movement, and the Institute for Strategic Dialogue. These organizations bring expert knowledge of complex issues like hate speech, radicalization, and terrorism that will help us better identify content that is being used to radicalize and recruit extremists. We will also regularly consult these experts as we update our policies to reflect new trends. And we’ll continue to add more organizations to our network of advisors over time.
Tougher standards: We’ll soon be applying tougher treatment to videos that aren’t illegal but have been flagged by users as potential violations of our policies on hate speech and violent extremism. If we find that these videos don’t violate our policies, but contain controversial religious or supremacist content, they will have some features removed. The videos will remain on YouTube behind an interstitial, won’t be recommended, won’t be monetized, and won’t have key features including comments, suggested videos, and likes. We’ll begin to roll this new treatment out to videos on desktop versions of YouTube in the coming weeks, and will bring it to mobile experiences soon thereafter. These new approaches entail significant new internal tools and processes, and will take time to fully implement.
Early intervention and expanding counter-extremism work: We’ve started rolling out features from Jigsaw’s Redirect Method to YouTube. When people search for sensitive keywords on YouTube, they will be redirected towards a playlist of curated YouTube videos that directly confront and debunk violent extremist messages. We also continue to amplify YouTube voices speaking out against hate and radicalization through our YouTube Creators for Change program. Just last week, the U.K. chapter of Creators for Change, Internet Citizens, hosted a two-day workshop for 13-18 year-olds to help them find a positive sense of belonging online and learn skills on how to participate safely and responsibly on the internet. We also pledged to expand the program’s reach to 20,000 more teens across the U.K.
And over the weekend, we hosted our latest Creators for Change workshop in Bandung, Indonesia, where creators teamed up with Indonesia’s Maarif Institute to teach young people about the importance of diversity, pluralism, and tolerance.
Altogether, we have taken significant steps over the last month in our fight against online terrorism. But this is not the end. We know there is always more work to be done. With the help of new machine learning technology, deep partnerships, ongoing collaborations with other companies through the Global Internet Forum, and our vigilant community we are confident we can continue to make progress against this ever-changing threat. We look forward to sharing more with you in the months ahead.
We wanted to give you an update on these commitments:
Better detection and faster removal driven by machine learning: We’ve always used a mix of technology and human review to address the ever-changing challenges around controversial content on YouTube. We recently began developing and implementing cutting-edge machine learning technology designed to help us identify and remove violent extremism and terrorism-related content in a scalable way. We have started rolling out these tools and we are already seeing some positive progress:
- Speed and efficiency: Our machine learning systems are faster and more effective than ever before. Over 75 percent of the videos we've removed for violent extremism over the past month were taken down before receiving a single human flag.
- Accuracy: The accuracy of our systems has improved dramatically due to our machine learning technology. While these tools aren’t perfect, and aren’t right for every setting, in many cases our systems have proven more accurate than humans at flagging videos that need to be removed.
- Scale: With over 400 hours of content uploaded to YouTube every minute, finding and taking action on violent extremist content poses a significant challenge. But over the past month, our initial use of machine learning has more than doubled both the number of videos we've removed for violent extremism, as well as the rate at which we’ve taken this kind of content down.
More experts: Of course, our systems are only as good as the the data they’re based on. Over the past weeks, we have begun working with more than 15 additional expert NGOs and institutions through our Trusted Flagger program, including the Anti-Defamation League, the No Hate Speech Movement, and the Institute for Strategic Dialogue. These organizations bring expert knowledge of complex issues like hate speech, radicalization, and terrorism that will help us better identify content that is being used to radicalize and recruit extremists. We will also regularly consult these experts as we update our policies to reflect new trends. And we’ll continue to add more organizations to our network of advisors over time.
Tougher standards: We’ll soon be applying tougher treatment to videos that aren’t illegal but have been flagged by users as potential violations of our policies on hate speech and violent extremism. If we find that these videos don’t violate our policies, but contain controversial religious or supremacist content, they will have some features removed. The videos will remain on YouTube behind an interstitial, won’t be recommended, won’t be monetized, and won’t have key features including comments, suggested videos, and likes. We’ll begin to roll this new treatment out to videos on desktop versions of YouTube in the coming weeks, and will bring it to mobile experiences soon thereafter. These new approaches entail significant new internal tools and processes, and will take time to fully implement.
Early intervention and expanding counter-extremism work: We’ve started rolling out features from Jigsaw’s Redirect Method to YouTube. When people search for sensitive keywords on YouTube, they will be redirected towards a playlist of curated YouTube videos that directly confront and debunk violent extremist messages. We also continue to amplify YouTube voices speaking out against hate and radicalization through our YouTube Creators for Change program. Just last week, the U.K. chapter of Creators for Change, Internet Citizens, hosted a two-day workshop for 13-18 year-olds to help them find a positive sense of belonging online and learn skills on how to participate safely and responsibly on the internet. We also pledged to expand the program’s reach to 20,000 more teens across the U.K.
And over the weekend, we hosted our latest Creators for Change workshop in Bandung, Indonesia, where creators teamed up with Indonesia’s Maarif Institute to teach young people about the importance of diversity, pluralism, and tolerance.
Altogether, we have taken significant steps over the last month in our fight against online terrorism. But this is not the end. We know there is always more work to be done. With the help of new machine learning technology, deep partnerships, ongoing collaborations with other companies through the Global Internet Forum, and our vigilant community we are confident we can continue to make progress against this ever-changing threat. We look forward to sharing more with you in the months ahead.