Facebook has admitted that it is developing new artificial intelligence algorithms by harvesting Instagram’s vast library of public images.
The company revealed at its F8 Developer Conference that it has been pulling vast numbers of Instagram pictures and their associated hashtags in an effort to improve machine picture identification skills.
“We rely almost entirely on hand-curated, human-labelled data sets,” Facebook’s chief technology officer Mike Schroepfer said onstage, as reported by The Verge.
“If a person hasn’t spend the time to label something specific in an image, even the most advanced computer vision systems won’t be able to identity it.”
The company claims this tech is being used primarily for moderation efforts and to aid its battle against both fake news and potential electoral interference.
Schroepfer added: “Until very recently we often had to rely on reactive reports. We had to wait for something bad to be spotted by someone and do something about it. This is why we are so focused on core AI research. We require new breakthroughs, and we require new technologies to solve problems all of us want to solve.”
The news asks further questions not just about Facebook’s competitive advantage in the tech market but also about user data privacy – a subject that could not be hotter for Facebook right now.
The images used are all public, and it would seem as if the data gathered is not being used for the purposes of identifying people or subsequently targeting them for ads or data sales.
At the same time, Instagram users may well be completely unaware their content is being used in this manner. Plus, while teaching a machine to tell the difference between a beagle and bulldog is a relatively innocent endeavour, it is arguably less so when such skills will likely lead to commercial gain for a company.
At a time when the words ‘Facebook’ and ‘security’ are so hotly conjoined, this could be a risky revelation.