We are used to living among electronic devices in our homeand the mobile phones that accompany us wherever we go, we also have a series of smart appliances in our home that do not stop collecting data, such as a smart fridge or even that Vacuum cleaner robot that makes life so easy for us.
And one of the most controversial stories regarding the processing of data collected by a device of this caliber, specifically by a robot vacuum cleaner, happened at the end of 2020, where some images captured by a well-known robot vacuum cleaner ended up in private Facebook forums in Venezuela. .
Specifically, the leaked private images were originally taken by a developing version of the robot vaccum cleaner Roomba J7 from iRobot, which were later sent to Scale AI, a startup that collects and tags images to train artificial intelligence.
However, many of these private photos made it to MIT Technology Review, which denounced the situation and found out how those images originally captured by the robot vacuum cleaner could have reached a private Facebook forum.
Following the leak of images, iRobot confirmed that these images were captured by its Roomba in 2020 coming from special development robots with hardware and software modifications that were never present as consumer products for purchase.
They clarify that these robots were delivered to collectors and paid employees who previously had to sign written agreements acknowledging that they were constantly sending a stream of data, including video, to the company for training purposes.
According to iRobot, these special devices were tagged with a green tag that read “video recording in progress” and that it was subsequently up to the data collectors to remove anything they deemed sensitive.
The reason for this type of procedure
And it is that this type of technology is based on machine learning, which makes use of a large amount of data, including voices, faces and even elements of our home, to train algorithms and thus recognize patterns.
This data is then used to build smarter robots. And for these data sets to be useful, certain individual humans, such as these paid data collectors, must view, categorize, label, and add context to many of these captured images.
iRobot, on the lips of James Bausmannnoted that the company “has taken all precautions to ensure that personal data is processed securely and in accordance with applicable law”. It clarifies that these images that the medium accessed “were shared in violation of a written order”, that is, from a non-disclosure agreement between iRobot and an image annotation service provider.
Subsequently, the CEO of iRobot, Colin AngleClarified that “iRobot has terminated its relationship with the service provider that leaked the images, and is actively investigating the matter and taking steps to help prevent a similar leak by any service provider in the future.”.
And it is that today this practice of sharing potentially confidential data to train algorithms is becoming widespread, some of them like the ones in this story where a single image traveled through a multitude of places until it finally reached those private forums on Facebook.