Facial recognition software for the unemployed rarely works

When it comes to face recognition, every nation acts individually. While there is still discussion in Europe as to the extent to which such systems should be approved in the future, facial identification is already common practice in China and is unfortunately often used to the detriment of many people. In the USA, however, each state decides for itself how it wants to proceed here.

An example from the land of unlimited possibilities now proves that face recognition systems are not error-free either. Many states now use such a system to approve or reject funds for the unemployed. This is to prevent abuse. However, new reports suggest that even authorized persons often do not receive any funds.

The ID.me system is used here, which approves or rejects an application based on official documents and biometric identification features. What one has apparently only inadequately thought about is the reaction to rejected applications when the person is actually entitled to benefits. Because these usually do not reach a human processor.

This is a problem because many Americans have lost their jobs precisely because of the coronavirus pandemic. If they receive money that has not been promised by the state, they slide into poverty and homelessness is a real danger. Even so, several states took the AI ​​route last year after scammers took advantage of the pandemic.

Overall, it was promised to reduce damage by 400 billion dollars, but between March and October 2020 it was only possible to save 5 billion dollars from fraudsters. This is a large sum compared to the risks that the system has brought, but it is hardly tenable. It remains to be seen how the story unfolds here.

Own opinion:

Does facial recognition really make sense and useful? If you look at the experiences of the past years, you can see that the dangers usually outweigh the actual benefits. In this country a pilot test has been started at the Berlin Südkreuz train station, but the results were not really satisfactory. One more reason to forbid the use of such systems in principle.