Click below nairabet banner for quick download
Almost every technology company is engaged in the development of facial recognition system. One of them is Amazon — its technology is called Rekognition and, according to experts from the Massachusetts Institute of technology, it is much inferior to its analogues. The fact is that in assessing its accuracy, it was found that the system often could not distinguish the sex of black people. This is the second complaint about Amazon’s abusive facial recognition work.
Face recognition system
Scientists tested the work of artificial intelligence Rekognition throughout 2018. The number of errors in determining the sex of black people reached 31% and this, according to researchers, a terrible result. For comparison, Microsoft facial recognition allows such errors only in 1.5% of cases.
Amazon believes that the claims of researchers are unfounded. She assured that in internal tests of the updated version of Rekognition it wasn’t noticed any mistakes. The researchers noticed that MIT employees did not name the level of accuracy at which the work of artificial intelligence Amazon would be considered “correct”.
According to CEO of deep learning Amazon Web Services Matt wood, they independently tested the system using a million face images from the Megaface database. During their testing, artificial intelligence did not make a single mistake. At the same time, the company’s representative announced that they are ready to listen to the results of third-party tests in order to improve their product.
This is not the first time Amazon’s artificial intelligence has been at the center of a scandal. In the summer of 2018, she passed 28 members of Congress for criminals, 38% of whom were black.
Other companies have also dealt with similar scandals. Fortunately, they learn from situations and constantly improve technology. In June, Microsoft expanded the amount of data used in facial recognition — in particular, the system began to pay more attention to gender, age and skin tone. This reduced the number of errors in sorting men and women up to 20 times.
Do you found this post helpful? Here is some action you might like to take. - Add comment below
- Share on social media
- Like our social pages for more updates
- Subscribe for notification by clicking the bell icon on the screen.