Meta Warns of Headwinds, FIFA Builds Offside Detector, and Other AI News

We aim to inform readers not only about developments in the bitcoin industry but also about adjacent technological fields—cybersecurity and the world of artificial intelligence (AI).

ForkLog AI has gathered the most important AI news from the past week.

Meta warns of headwinds ahead

The company Meta warned employees of ‘tough times’. It also outlined priorities and challenges for the future.

Product chief Chris Cox said Meta faced financial headwinds stemming from changes to privacy policy and macroeconomic pressure. Employees must ‘perform flawlessly’ without expectations of a huge influx of new engineers and budgets, he added.

The executive highlighted six areas in which the tech giant must increase investment:

• metaverse;
• artificial intelligence;
• messaging;
• growth of Reels;
• monetization;
• compliance with new privacy requirements.

The company also needs to quintuple the number of GPUs in data centers to compete with TikTok, Cox said.

Italy will allocate €45 million to AI and blockchain developers

The Italian government will allocate €45 million to subsidize projects related to AI, IoT and blockchain. Applications for funding will open on 21 September 2022.

Officials noted that grants could go to developers of socially significant projects such as education, agriculture and health.

A new investigation has been opened into Tesla

The U.S. National Highway Traffic Safety Administration (NHTSA) has opened an investigation into two fatal Tesla crashes.

One incident involved the deaths of two Florida residents on July 6, 2022. According to the highway patrol, the vehicle collided with the rear of a stationary tractor-trailer parked in a truck stop. The driver and passenger died.

A highway patrol spokesperson noted that it is not yet known whether Autopilot was engaged at the moment of impact.

FIFA will use AI to detect offsides

The International Federation of Football (FIFA) will deploy AI to determine offsides at the 2022 World Cup in Qatar.

According to representatives, the system includes a sensor in the ball and 12 cameras on the stadium roof. Also, thanks to machine learning, the video sensors will track 29 points on a player’s body.

The collected data will allow the system to detect an offside within a few seconds. In case of violation, the program will relay information to FIFA officials who will alert the on-field referee.

The organization stressed that the system is intended to help make faster and more accurate decisions, not replace humans.

Elon Musk: General AI needs regulation

The American entrepreneur Elon Musk believes that artificial general intelligence should be regulated.

In an interview for the Getting Stoned podcast, the entrepreneur said that if current progress continues, humanity will soon be able to create digital intelligence that surpasses human in every respect.

He added that general AI could pose risks to public safety. Therefore the technology should be subject to regulatory oversight, Musk believes.

Korean police to expand the role of robots and algorithms in law enforcement

By 2050, South Korean law enforcement plans to deploy robotic dogs, unmanned patrols and AI.

According to the National Police Agency (NPA), autonomous devices will help keep streets in order. AI technologies will enable greater mobility for law enforcement, experts say.

The NPA added that the innovations will help agencies adapt to a changing world, with the main aim of reducing police reliance on humans.

Scientists create AI tool to improve mental health

Deep Longevity, in collaboration with Harvard Medical School, has developed AI that improves mental health.

The algorithm rests on two models. The first is a set of neural networks predicting respondents’ chronological age and psychological well-being 10 years ahead from survey data. This model maps trajectories of the human mind as it ages.

The second is a self-organizing map used as the basis for a mental-health improvement recommendation system. The algorithm clusters respondents by the probability of developing depression and identifies the shortest path to the cluster of mental resilience.

Deep Longevity also released a free web app allowing users to take a psychological test and obtain a report with ideas for improving long-term mental well-being.

GPT-3 writes a scientific paper about itself

The GPT-3 text generator wrote an academic paper about itself. Researchers submitted the “paper” for peer review to a scientific journal.

Swedish researcher Almira Osmanović Tunström asked the algorithm to write a 500-word scientific article about GPT-3, and to add references and quotes. According to her, after two hours the system completed the document, placing verified quotes in the right places and in the correct context.

Tunström suggested that the experiment could change the funding system of the academic community depending on publication counts. In her words, the AI assistant could generate a new article every day.

Scientists present algorithm for generating three-dimensional images

Researchers from Stanford developed an AI model EG3D that generates random high-resolution 3D images of faces and other objects.

Engineers added an extra component to the generative adversarial network that turns 2D images into 3D.

To edit photos created by EG3D, another team of researchers developed a machine learning model called GiraffeHD. According to engineers, both algorithms demonstrated high speed even on low-end laptops.

Enthusiast creates an AI controller that fires using voice commands

The American streamer Dylan Beck, known as Rudeism, developed an AI controller for Apex Legends.

The device fires with voice commands \”br-r-r-r\” and \”paaf-paaf\”, while aiming with a special glove.

As a basis for the controller, the modder used an Arduino single-board computer with a microphone and a motion sensor mounted on the arm.

The week’s most notable AI deals

Between July 3 and 9, 2022, AI startups attracted over $90 million. Here are the most interesting deals.

• Klang Games attracted $41 million to create a virtual world inhabited by AI-powered beings.
• The Celus startup, using AI to automate circuit-board design, raised $25.6 million.
• AI chipmaker Rebellions raised $22.8 million.

Also on ForkLog:

• In the U.S. will build a supercomputer ten times more powerful than current systems.
• DeepMind trained AI “fairly” to distribute wealth.
• In China, AI for assessing loyalty of Communist Party members was created.
• The EU passed bills, regulating the algorithms of tech giants.
• Meta unveiled an AI translator with 200 languages.
• Researchers created AI that determines salary based on job descriptions.

Subscribe to ForkLog AI news on Telegram: ForkLog AI — all the news from the AI world!