How were intimate photos taken by a robot vacuum published on social media?
A woman sitting on her toilet with her shorts down to mid-thigh, a young boy lying on his stomach in a hallway, a playful dog, untidy interiors, others much better organized… Thousands of photos taken by robots The vacuum cleaners were published by errors on social networks and especially on Facebook.
The MIT Technology Review site uncovered this case, which dates back to 2020 when employees of a Venezuelan provider of a major brand of robot vacuum cleaners published a series of images on online forums to discuss the data mentioned by these so-called intelligent objects.
In the middle of the photos, domestic scenes but also other more intimate ones which should never have ended up on the internet.
To improve artificial intelligence
According to MIT Technology Review, the photos were not taken without the knowledge of the robot’s owners. They had given their consent and were even paid to participate in a vast program to improve the system.
As such, the colossal data recorded by the vacuum robots were regularly presented to these suppliers located around the world in order to improve the artificial intelligence of the robot.
While typically delivered to the cloud, this data is supposed to be ultra-secure and never leave the internal network. Nothing explains today why they were broadcast on Facebook.
They knew they were being filmed
The case is causing a lot of noise and is taken very seriously by iRobot, the world market leader, recently acquired by Amazon for 1.7 billion dollars. In a press release, iRobot recalls, however, that participants in their program are warned of the recording of a video by a green light located on the robot and that they have the ability to “delete any content deemed sensitive captured in the space where the robot works, including images of children”.
In other words, iRobot considers that the people filmed have tacitly agreed to let their Roombas monitor them.