Site icon

Ear recognition, US TikTok ban and other AI news


We strive to inform readers not only about events in the bitcoin industry, but also to talk about what is happening in related technological areas – cybersecurity and in the world of artificial intelligence (AI).

Cryplogger AI collected the most important news from the world of AI over the past week.

  • The US Department of Commerce has placed 36 Chinese companies on the sanctions list.
  • Google CEO Sundar Pichai called AI chatbots immature.
  • US congressmen proposed to ban TikTok.
  • Australian artists have accused AI self-portrait app Lensa AI of stealing their work.
  • Scientists have created a system for identifying people by their ears.
  • TikTok algorithms promoted Billie Eilish’s pornographic deepfakes in the For You tab.
  • The most important AI deals of the last week.

US blacklists over 30 Chinese companies

US Department of Commerce added 36 Chinese enterprises are on the list of companies that cannot import certain American technologies without a license.

According to the agency, 30 organizations participated in the modernization of the military-industrial complex of China. Of these, 21 companies used AI chips in their activities.

One organization came under restrictions for allowing human rights violations, and another for supporting the production of Iranian drones.

The remaining four companies were added to the list due to “negative impact” on US national security or foreign policy.

Sundar Pichai called chatbots with AI immature

Google has no plans to launch a ChatGPT competitor yet due to “reputational risks”.

According to CNBCat a general meeting, staff asked management if the rollout of the OpenAI smart chatbot was a “missed opportunity” for the tech giant.

Alphabet CEO Sundar Pichai and head of Google’s AI division, Jeff Dean, responded that the corporation’s language models are no less effective than competitor solutions. The tech giant needs to move “more conservatively” than a small startup, they added.

Pichai and Dean called AI chatbot technology “immature” and pointed to issues of bias, toxicity, and a tendency to “make up” information.

They also said that Google has “many” new AI language features planned for 2023, an area that requires the company to “be bold and responsible.”

TikTok proposed to be banned in the US

American congressmen suggested ban TikTok over fears of Chinese spying.

Members of the US Senate and House of Representatives have developed a bipartisan bill, aimed at “blocking any transactions” of the service. It will also ban the use of all social media companies located or under the “significant influence” of Venezuela, Iran, China, Cuba, North Korea and the Russian Federation.

Apps like this have the potential to leak US user data to repressive governments, censor news and manipulate the public, lawmakers say.

TikTok representatives expressed concern that the congressmen submitted the document before the completion of the state security review.

December 15 US Senate accepted a bill banning the use of the platform in federal agencies. Now the document must be approved by the House of Representatives.

Media: Lensa AI popularized AI applications in the App Store

The success of the Lensa AI service has sparked a rise in the popularity of art AI apps on the App Store.

According to TechCrunch on December 12, 2022, the top three lines of the American marketplace were occupied by photo editors with artificial intelligence. In total, eight such applications were in the top 100 by the largest number of downloads in the App Store.

The first place is still occupied by the Lensa AI service, which was installed by 12.6 million users around the world in 11 days of December. This is 600% more than the figure for November 20-30.

The second and third places were taken by AI Art: AI Image Generator and Dawn – AI Avatars, which were downloaded 1.7 million times. At the end of November, the number of downloads was 71,000 and 28,000, respectively.

Artists accuse AI app of stealing their work

Australian artists accused Lensa AI self-portrait app in stealing their work.

The tool is based on the Stable Diffusion algorithm trained using the LAION-5B online image database. According to the artists, the dataset includes many works taken without permission. The app copies different styles, down to strokes, colors and compositions, they added.

The artists said they had not received any compensation and it “looks like a violation”.

The artists called for tougher copyright laws to “keep up with the evolution of AI art.”

The company responded that Lensa AI trains to generate content in the same way as a person – by studying various artistic styles. Terms like “counterfeiting”, “theft of works of art” or “illegal use” cannot be applied to this process, the developers added.

Google Introduced AI Robot Training Algorithm

Google Explorers submitted the RT-1 transformer model, which allows robots to perform more than 700 tasks with a 97% success rate.

The experts trained the algorithm on data that included 130,000 videos of 13 Alphabet Everyday Robots devices performing various actions. Due to this, the model learned to determine the next most likely movement when performing operations by the robot.
Robot learning process. Data: Google.

The approach allows the system to generalize the knowledge gained during training, and apply experience to solve new problems. For example, trained on many videos about how to pick a banana, the machine will be able to identify and grab the fruit even in an unfamiliar environment.

Scientists have created an ear recognition system

University of Georgia researchers developed a system for identifying people by their ears. The accuracy of the algorithm is about 99% depending on the data set and model.

The software works in a similar way to facial recognition technology. The user needs to provide the algorithm with several images of ears so that it can recognize their owner.

Ear recognition algorithm. Data: University of Georgia.

During tests on two existing datasets, the system showed an accuracy of 97.25% and 75.11%, respectively. The researchers also found that the software is able to process images with blurs, noise, and different variations in brightness or contrast.

According to scientists, the technology can be used to improve existing security systems.

OPPO introduced Robops

The Chinese manufacturer of equipment OPPO has shown a robo-companion QRIC.

The device can walk, jump, somersault, carry objects and recognize people. The robot supports voice and remote control.

The company plans to reveal the technical details of QRIC at the OPPO INNO Day 2022 conference on December 14, 2022.

Robo-barist implemented in American offices

Startup Artly involved robobarist in seven offices in the USA.

The manipulator brews espresso, pours milk, froths and brings everything together to prepare a cappuccino. The device also draws latte art on the drink.
Robobarista makes coffee. Data: Artly.

In each office, the developers installed different robots. They differ in name, “personality”, “gender”, voice and accent. The devices also use unique approaches to making coffee, as they were trained by different baristas.

TikTok distributed Billie Eilish deepfake porn for four days

TikTok Algorithms promoted pornographic deepfakes by singer Billie Eilish in the For You tab.

The video included several photos of the performer’s face along with an artificially “sexualized” body. The video received 11 million views in four days before being removed by TikTok for violating community guidelines.

Tracey Harwood, a professor of digital culture at De Montfort University, said the images looked like they were using some kind of AI application.

The account that owns the video has over 76,000 followers and is linked to an Instagram account with over 2,000 followers. There, the author also shared a video and offered to buy “uncensored images” for $20.

The most important AI deals of the last week

From December 11 to December 17, 2022, AI startups raised over $620 million. Here are the most interesting transactions.

  • Creators of Dataiku, a platform for designing, deploying and managing AI applications attracted $200 million
  • Zappi received $170 million to expand consumer analytics service.
  • Shield AI collected $60 million to develop software and hardware for military drones.
  • The creators of the platform for automating accounting processes got $52 million
  • Beamery attracted $50 million for the development of a service for personnel management.
  • Sana Labs collected $34 million to support a platform for monitoring work information and transforming it into a learning resource.
  • AI hardware accelerator manufacturer EnCharge AI received $21.7 million
  • LexCheck collected $17 million to expand the contract analysis platform.
  • AI security company Protect AI attracted $13.5 million
  • Anima received $3 million to support an NFT project with AR pets.

Also on Cryplogger:

Subscribe to Cryplogger news in Telegram: Cryplogger AI – all the news from the world of AI!

Found a mistake in the text? Select it and press CTRL+ENTER

Exit mobile version