Tech Giants Unite to Take Down Terrorist Posts

Some of the biggest Internet firms say they'll work together to remove images and videos that promote the message of terrorists. The announcement comes in the same week that European officials have criticized companies for failing to live up to a promise to tackle hate-based speech online.

The agreement involves Facebook, Microsoft, Twitter and YouTube. They'll build a joint database of images and videos that they have removed from their site after human moderators have flagged them as being either violent imagery produced by terrorists, or imagery specifically designed to recruit people to terrorism.

Once an image or video is on the database, the other companies will be able to automatically search for it among posts by users and, if they choose, remove it or block it from being uploaded in the first place. It's a similar approach to existing efforts to fight obscene images of children being shared online.

Database Spots Edited Pics

The database works by producing hashes, also dubbed "digital fingerprints." The hash is made of code that identifies key characteristics of the image. That makes it easier to automatically find versions of the offending images or video that have been edited, for example by cropping them differently, in an attempt to beat automated filters.

Although the companies will share the database, they say each will maintain their own rules and standards governing which images and videos to remove. They'll also continue to operate individual policies about how to comply with government requests to take down imagery and whether or not they make the details of such requests public. (Source: fb.com)

Tech Firms Too Slow On Hate Speech

Meanwhile the European Commission says big tech firms have not met a promise made earlier this year to take down all text-based posts containing illegal hate speech within 24 hours of first receiving a report about the post. Instead the Commission says its research shows this is currently happening just 40 percent of the time.

The figures are based on research made by 12 agencies across nine countries. The research also found that only 28 percent of reports led to a takedown at all. That's likely because the agreement only covers hate speech that is illegal in the relevant country rather than merely something that causes offense. (Source: europa.eu)

What's Your Opinion?

Is the database a smart way of tackling violent or pro-terrorist imagery? Will automated filtering become more important as the amount of content online continues to grow? Is it reasonable to expect illegal text-based posts to be taken down within 24 hours of a report?