Robot vacuum cleaners took pictures inside apartments and sent pictures to third parties

In the West, a scandal erupted around the fact that the robot vacuum cleaner not only took pictures inside apartments, but also sent pictures to third-party companies. Subsequently, screenshots of these photos appeared on Facebook. About it пишет edition of the MIT Technology Review.

All of the images were taken with an iRobot Roomba J7 series robot vacuum and then sent to Scale AI, a startup that hires employees around the world to label audio, photo and video data used to train artificial intelligence. As a result, private pictures of people were transferred to the company, including a woman sitting on the toilet with her pants down. There were also shots of children, homes around the world, and people busy with their own business or animals.

iRobot is the world's largest supplier of robotic vacuum cleaners. The company confirmed that these images were taken by Roomba robots in 2020, but said that the technology was not present on consumer products for purchase - only as part of development versions of the vacuum cleaner. The company also said that the people present in the photographs gave their consent to send data to the company for artificial intelligence training, and the devices themselves were marked with a bright “Recording” sticker.

The 15 photos leaked online are just a tiny fraction of the total. According to iRobot, more than 2 million such images have been made.

At the same time, the company refused to give an opportunity to get acquainted with the agreement on this, agreed and did not give the opportunity to communicate with any of the employees participating in the testing.

The situation is serious in that when using different devices consumers regularly agree to the monitoring of various data programs, devices - from iPhone to washing machines.

This practice has become more common in recent years as data-intensive artificial intelligence is being integrated into a range of new products and services. And the technology itself is mainly based on machine learning, which uses large amounts of data, including people's voices, faces, houses and other personal information. In this manner algorithms are trained to recognize patterns.

However, the most useful datasets are the most realistic data, which makes data obtained from, for example, a human dwelling the most valuable. And often users simply agree to the user agreement without even reading it in order to continue using the product.

In addition, privacy policies can be quite vague and give companies a lot of leeway in how they share and analyze the information they receive from users.

The material notes that the data received from robotic vacuum cleaners can be especially useful because they have powerful equipment and sensors, and they can independently move around a person’s house. This data is then used to build smarter robots, whose work can go far beyond the normal operation of a vacuum cleaner.

Note what happens when applications that do not need many functions to work request additional permissions that are not included in their functionality.

In addition, cybersecurity experts warn that if, for example, you have given the Camera app permission to take photos and record videos, it will be able to do it on its own at any time without warning the owner of the phone. The same goes for recording the sound of a microphone, which can record everything that happens near the smartphone.

The trick of this work of applications is that often without the provision of all permissions they cannot be used, that is, you have to agree at your own peril and risk. On the other hand, if you don't give the Camera app access to your camera, it won't work properly. Therefore, it is better to grant the necessary permissions only to applications that work by default or that you are sure about.

We also wrote about how a talking robot was brought to the Verkhovna Rada and asked him whose Crimea, and told that Ukraine arrested Belarusian equipment for 200 million hryvnias.

Subscribe to our news in Google News
Follow to telegram channel Klymenko Time
Follow to analytical telegram channel Klymenko Time

Did you like the material? Give him 5 stars. This is important for editors.