Now Google also joins the ranks of those tech companies that warn against unwanted consequences of facial recognition. The company said in a blog post that it would first refrain to release its own software for facial recognition to the public.

The technology must first be carefully scrutinized so that "its commitment complies with our principles and values, and avoids abuses and harmful consequences," writes Kent Walker, Google's Legal Department Director. With a sideless to the competition, Walker announces that facial recognition is initially not commercially available. Unlike other companies, they wanted to "first clarify important technical and political issues".

Civil rights activists of the American Civil Liberties Union (ACLU) refer to the announcement as a "strong first step". Thus, the group has shown that he has a moral compass. "We will renew our appeal to Amazon and Microsoft not to continue offering their dangerous face monitoring to the government."

Amazon's surveillance patent application

These two companies are also the swipe of Google. Because both companies have released their face recognition software for customers to earn money with it. Although Microsoft recently warned about the risks of technology such as facial recognition, which is based on machine learning, but offers the corresponding software long ago: Customers can access the Azure interface and face recognition at ridiculous prices.

Amazon, in turn, is accused of selling the technology as a service to, among others, the controversial US Immigration Service (ICE), which is said to have separated children from their parents on the Mexican border.

According to a patent application Amazon is even thinking about a more extensive privatization of video surveillance: The ACLU had published Wednesday the plans of a networked doorbell with camera. Accordingly, Amazon wants to let all the passers-by film passing by the front door. The faces are to be matched with a database so that both the homeowner and the police suspects are reported.

According to the ACLU, this could lead to Amazon assembling "all parts of a surveillance network that ranges from the government to the front door" with such devices and the servers they provide to law enforcement.

REUTERS

FILE PHOTO: A Google logo is seen at the companys headquarters at Mountain View, California, US, November 1, 2018. REUTERS / Stephen Lam - RC1D55999C70 / File Photo

But Google has not necessarily acted in the past on the values ​​that the company now propagates. So let go of Google a contract with the US military only on pressure of the own coworkers. Google had automatically evaluated the recording of drones with self-learning software (called Artificial Intelligence or AI for short).

No opt-out possible

It is not possible without rules, scientists from New York University at the AI ​​Now Institute agree. In their annual report (PDF), researchers demand that such technology be more tightly regulated by law and that companies internally monitor ethical development. The aim is to prevent a surveillance state and discrimination against women and non-whites in particular.

In principle, the researchers recommend that everyone should have the chance to refuse automatic detection and analysis. But due to the system, all faces are always filmed first. Only in the next step could a software recognize that a person does not want to be analyzed.