FACEBOOK has secretly changed the privacy settings of users outside the US by activating new technology that automatically recognises people in photographs.

The social networking company, which has previously come under scrutiny for its privacy practices, admitted today that it should have been “more clear” with users before the launch of the feature.

The new facial recognition technology analyses photographs to detect if a person’s face belongs to a fellow Facebook user and then encourages their online friends to “tag” them – the practice of assigning a name to a person who appears in a photograph.

However, Facebook does not give users the option to avoid being tagged in an image, and they must manually “untag” themselves if they do not want their name linked to a photograph.

The feature, “Tag Suggestions”, went live in the US in December and was rolled out across the rest of the world in the past few days without warning, according to internet security firm Sophos.

“Yet again, it feels like Facebook is eroding the online privacy of its users by stealth,” Graham Cluley, a senior technology analyst at Sophos, wrote on the company’s blog.

Facebook said, “We should have been more clear with people during the roll-out process when this became available to them”.

The company said that the new feature only suggests tagging people who are already friends with the user who posts the image, so strangers cannot be identified on other people’s photographs.

It added that its 500 million users can opt out of the feature by customizing their privacy settings.

Facebook has a history of automatically updating its users’ settings to activate new features, with privacy advocates arguing that users should have to “opt-in” to new technology that could affect their privacy.