US company Clearview AI is building a database that will allow face recognition of almost all people on Earth. Photos and information are collected from social networks and other Internet resources without the permission of users. Clearview AI representatives told investors that by the end of 2022 they will create a "face index" of 100 billion photos - so there will be 14 photos for each of the 7 billion people on the planet. They are used for a surveillance system that helps law enforcement agencies around the world find people to arrest or conduct criminal investigations. This is written by The Washington Post, referring to the presentation of the developers. It is noted that since the beginning of 2020, the database has grown from 3 billion to 10 billion images, while the system processes about 1.5 billion photos per month. Clearview's technology was said to be larger and more comprehensive than the systems used in China, as the "face database" is linked to "metadata from publicly available sources." Clearview customers can upload a photo to search for a match in the company's face database, including online accounts. The company said its "face index" is already 11 times larger than any government or non-government agency's face database. WP emphasizes that Clearview AI is funded primarily by the US budget and government contracts, and its products are controversial. The presentation was arranged to ask investors for additional funds. As the company noted, investments in the amount of $50 million will allow to supplement the database, expand the team of employees and pay politicians more for lobbying the necessary decisions at the state level. The presentation shows that Clearview AI considers government contracts as a small part of potential revenue. The company plans to sell its system not only to law enforcement, but also to banks, retailers and e-commerce businesses. “If they're only selling their product for general commercial use, it's just mass surveillance on a massive scale. It is not intended for the most extreme cases, as they promised earlier, ”commented Jack Poulson, head of the Tech Inquiry research team. The company also said that it doesn't want to be limited to face recognition - Clearview AI is working on several new technologies that will allow people to be found by photos and gait, as well as to read fingerprints from a distance. Airbnb, Lyft and Uber are already interested in the technology, Clearview AI CEO Hoan Ton-Tat told Motherboard. However, the companies denied this information. According to WP, while US law does not regulate the use of facial recognition, although some cities and states have adopted local bans or restrictions. Major tech giants, including Amazon, Google, IBM and Microsoft, have limited or stopped sales of such systems, saying they are concerned about the risks and don't want to launch them until the systems are finalized. Facebook, Google, Twitter and YouTube have demanded that Clearview AI stop collecting photos from their sites and delete everything it has accumulated. Clearview itself refuses to make concessions, citing the First Amendment to the US Constitution, which provides for freedom of speech and the press. Some government officials also spoke out against Clearview AI, in particular Senators Ron Wyden and Rand Paul proposed a bill banning its funding from the state budget. Now the company is dealing with a wave of lawsuits and restrictions, and not only at home. In Sweden, the authorities fined the local police department for using the Clearview AI system in 2021. The company has also faced a class action lawsuit in federal court in Canada, government investigations in Canada, Sweden and the UK, and data protection complaints in France, Greece, Italy and the UK. The governments of Australia and the UK have ordered developers to delete the data of their citizens. The modern level of microprocessors makes it possible to almost completely copy all the currently discovered functions of the human neuron. As it turned out, human brain structures are often extremely redundant. That is why a rather complex face recognition function can today be implemented within a very modest neural network.