Facebook says it removes two-thirds of IP-infringing content

Social media giant Facebook says it removed more than 462,000 pieces of content related to counterfeits in the first six months of the year, with another 359,000 taken down from Instagram.

Figures from the company’s latest Transparency report suggests that is equivalent to a removal rate of between 62 and 72 per cent over the period on its main social networking site, with a similar takedown rate on social networking and picture sharing app Instagram.

It received 62,300 reports about counterfeit goods on Facebook in the first half, with another 39,200 reports on Instagram. It also took down 3,234,393 pieces of content based on 568,836 reports of copyright infringement and 255,222 pieces of content based on 96,501 trademark reports.

Facebook has come under fire from groups who say that its Marketplace store has turned into a black market for counterfeit or ‘replica’ goods.

A survey of 1,500 US consumers by Red Points earlier this year – focusing just on beauty products – suggested that 17 per cent of respondents had bought a gift that turned out to be a counterfeit, and 49 per cent of those purchases were made on either Facebook or Instagram.

Almost three quarters (73 per cent) of people polled said they regularly see social media content selling products they believe are counterfeit.

In a blog post, Facebook insists brand owners can directly report Facebook or Instagram content that is suspected to infringe IP, whether it is “a profile, page or group, or is an ad or Marketplace listing.”

“Content is typically removed within a day or less after we receive a rights holder’s report, and often faster – within a few hours or even minutes,” it adds.

That’s a reactive way to tackle the problem of course, and Facebook claims it is also developing tools based on machine learning and artificial intelligence to automatically remove IP-infringing content.

It says it is using automation to review Marketplace posts and ads before they go live, and is blocking content that may be selling counterfeits based on signals such as brand names, logos, keywords, prices, discounts and other suspicious indicators.