Skip to Content
Silicon Valley

Facebook says it’s getting better at weeding out child sex abuse images

November 14, 2019
Facebook
FacebookAP

The news: Facebook has revealed a huge jump in images of child sex abuse uploaded to the website: 11.6 million pieces of content from July to September this year, versus 6.9 million from April to June. However, it proactively detects and removes over 99% of all such images, and the large increase is due to better detection, it said in its latest transparency report.

The fake problem: The report both illustrates the vast scale of the task Facebook faces in acting on its own policies against hate speech, harassment, and child abuse and how sophisticated Facebook’s automated systems have become at picking up this sort of content. However, plenty still slips through the net. Facebook has revealed the scale of the problem of fake accounts: it removed 3.2 billion from its service from April to September this year, and estimates that about 5% of its 2.45 billion user accounts are fake. These fake profiles are used for everything from political manipulation to lucrative scams.

Hand it over: The report also shows a spike in requests for Facebook user data from governments around the world, with the US leading the pack, followed by India, the UK, Germany, and France. There were 50,741 demands for user data from US authorities in the last quarter, with 88% of them resulting in data being handed over. Two-thirds came with a gag order, meaning the subject cannot find out if the data has been accessed.

Instagram joins the party: This is the first time Facebook has reported content enforcement figures for Instagram, which it owns. It shows Instagram lags significantly behind at detecting banned content. For example, while all but 3% of suicide and self-harm content is taken down from Facebook without reports from users, Instagram users are still responsible for reporting 21% of the same type of content before it is removed. Instagram’s head of product, Vishal Shah, told reporters this reflects the fact Facebook has better-developed automated systems.

Sign up here to our daily newsletter The Download to get your dose of the latest must-read news from the world of emerging tech.

Keep Reading

Most Popular

Large language models can do jaw-dropping things. But nobody knows exactly why.

And that's a problem. Figuring it out is one of the biggest scientific puzzles of our time and a crucial step towards controlling more powerful future models.

How scientists traced a mysterious covid case back to six toilets

When wastewater surveillance turns into a hunt for a single infected individual, the ethics get tricky.

The problem with plug-in hybrids? Their drivers.

Plug-in hybrids are often sold as a transition to EVs, but new data from Europe shows we’re still underestimating the emissions they produce.

Google DeepMind’s new generative model makes Super Mario–like games from scratch

Genie learns how to control games by watching hours and hours of video. It could help train next-gen robots too.

Stay connected

Illustration by Rose Wong

Get the latest updates from
MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

Thank you for submitting your email!

Explore more newsletters

It looks like something went wrong.

We’re having trouble saving your preferences. Try refreshing this page and updating them one more time. If you continue to get this message, reach out to us at customer-service@technologyreview.com with a list of newsletters you’d like to receive.