We aim to inform readers not only about developments in the Bitcoin industry but also about what is happening in related tech fields—cybersecurity, and now in the world of artificial intelligence (AI).
ForkLog has gathered the week's most important AI news.
- By the end of 2021, four regions in Russia will launch robotaxi services.
- Yandex will deploy autonomous delivery robots on American university campuses.
- ARM's chief executive says that the Nvidia deal will benefit the semiconductor and AI sectors.
- Copilot reproduced entire fragments of code protected by copyright.
- In a year and a half, the number of AI companies in China almost quadrupled.
- About half of Russians say they trust artificial intelligence.
- A Belgian artist created an algorithm to recognise politicians who use smartphones during government meetings.
Commercial robotaxi services to launch in Russia by year-end
In four regions of Russia, an autonomous taxi will appear. said Deputy Minister of Economic Development Vladislav Fedulov.
He noted that the project became possible thanks to the passage of a law on experimental legal regimes. The first robotaxi will appear in 2021 in Moscow, Moscow region, Krasnodar Krai and Tatarstan.
The document also allows piloting technologies such as artificial intelligence, cargo transport using drones, telemedicine, and analytics of big medical and economic data.
Elon Musk acknowledges the challenges in building autonomous vehicles
Elon Musk did not anticipate that creating a self-driving car would be so difficult. The entrepreneur wrote on Twitter in response to questions about the release of the ninth beta version of the Full Self-Driving (FSD) autopilot.
Haha, FSD 9 beta is shipping soon, I swear!
Generalized self-driving is a hard problem, as it requires solving a large part of real-world AI. Didn’t expect it to be so hard, but the difficulty is obvious in retrospect.
Nothing has more degrees of freedom than reality.
— Elon Musk (@elonmusk) July 3, 2021
He promised that the new FSD version would be available “very soon.” Previously, the founder said the autopilot update would arrive in mid-May, later moved to June.
Yandex to launch autonomous delivery on US university campuses
Yandex and the American startup GrubHub will launch delivery of food using the Rover robots on 250 US college campuses. Robo couriers will appear on campuses in autumn 2021.
Autonomous couriers can move along sidewalks around the clock and in any weather, the company said. They can also navigate controlled and uncontrolled pedestrian crossings.
«Usually campuses cover large areas with many pedestrian zones. “Rovers” move along sidewalks and pedestrian paths and will be able to deliver to places where cars cannot reach», said the company.
Since 2017, Yandex's robots have logged more than 7 million autonomous miles. In this metric, the company trails only Google's Waymo, whose cars have covered more than 20 million miles in 12 years.
ARM chief backs merger with Nvidia
ARM's chief says the Nvidia merger would turn the United Kingdom into a global AI hub. The company said that the headquarters will stay in the United Kingdom, and Nvidia will invest in British talent. Segars added that the deal would open access to even more innovative solutions for the entire semiconductor industry.
«Together with Nvidia we will invest in the creation of new markets, not squeeze customers out of existing ones», said Simon Segars.
In April, the British government intervened in the multibillion-dollar deal between Nvidia and ARM on national security grounds. Regulators are due to publish a final report on the merger by the end of July 2021.
Copilot suspected of copying copyrighted code fragments
The AI-powered code-autocompletion tool GitHub Copilot lets users bypass license restrictions of the GNU General Public License by citing open-source projects.
I don’t want to say anything but that’s not the right license Mr Copilot. pic.twitter.com/hs8JRVQ7xJ
— Armin Ronacher (@mitsuhiko) July 2, 2021
According to the Twitter user mitsuhiko, the tool can reproduce verbatim a well-known function for fast inverse square root from Quake III Arena, the code base of which is licensed under GPL 2.0.
Other developers noted that such “code laundering” would allow bypassing the original copyrights.
On the official Copilot page at Copilot’s FAQ, it states that code created with AI is unique and owned by the user. The algorithm can reproduce code from training data verbatim only in 0.1% of cases.
China sees rapid growth in AI companies
Researchers found a sharp rise in the number of AI-related companies in 2020.
Today there are 439,000 AI-related companies registered in China. Of these, 172,000 opened in 2020, and 153,000 in the first half of 2021.
Nominal capital for 39% of firms ranges from 1 to 5 million yuan ($155,000–$775,000). Capital above 5 million yuan accounts for 37% of organisations, while the rest have less than 1 million yuan.
Analysts link the surge to increased state investment and the country's growing scientific prowess.
Almost half of Russians trust artificial intelligence
About 48% of Russians trust AI. Meanwhile 42% are skeptical, and 10% declined to answer. The All-Russian Public Opinion Centre (VTsIOM) questioned 1,600 residents aged 18 and over.
A third of Russians fear human replacement by algorithms at work. 67% would use the technology to work less while maintaining their current income.
94% of citizens believe that AI in medicine should present options, with the final decision remaining with a human.
The majority of Russians think the state should help foster AI development. The idea is endorsed by 79% of respondents.
44% of Russians believe AI in state and municipal governance will lead to fairer decisions.
Spanish researchers develop AI glasses for the visually impaired
A Spanish engineer and his wife, a doctor, developed a device with AI that helps their son with vision impairment navigate the surrounding world.
The device resembles a VR headset, equipped with cameras. The algorithms process the video feed in real time and create a three-dimensional image onto which text, graphics and video are overlaid on real-world images. The device will also alert the user to an object approaching that blocks the way.
To bring the glasses to market, the couple founded Biel Glasses, raised $1 million, and assembled a team of engineers and doctors. They have already obtained permission to operate the product in the European Union.
The device will go on sale at the end of 2021 in Denmark and Spain for $5,800. The developers say each unit will be tailored to the needs of a specific user.
Russia develops robo-tuna for saving fish from ecological disasters
Scientists at the neurotechnology and machine learning centre of the Baltic Federal University named after Kant (BFU) developed a model underwater robot-tuna to control the movement of fish schools.
The biomimetic robot will be able to replicate the movement, navigation and interaction with other animals in a natural environment.
Scientific researcher Viktor Kazantsev said the robo-tuna as a leader or predator could help move fish away from the epicentre of an ecological disaster, for example during an oil spill at sea.
He added that the team still needs to think through all the details of the device, from the chassis to creating AI that can make navigation decisions underwater, where GPS is unavailable.
Belgian AI that teases officials using smartphones
A Belgian artist Dries De Poorter developed a computer-vision algorithm that recognises politicians who use phones at work.
Every meeting of the flemish government in Belgium is live streamed on a YouTube channel.
Software is searching for phones and tries to identify a distracted politician. This with the help of AI. The results are then posted to social with the politician tagged. By @driesdepoorter pic.twitter.com/QUQHiYx2HI— The Flemish Scrollers (@FlemishScroller) July 5, 2021
The algorithms process each live broadcast from the Flemish government's YouTube channel. If a phone is detected in a person's hand, the program identifies the official using facial recognition. Then the video featuring the distracted minister is posted to the project's Twitter and Instagram accounts, tagging the politician.
AI cheats with auto-aim appear on gaming consoles
AI bots have been trained to cheat in console video games. The corresponding video was posted by the Twitter account Anti-Cheat Police Department.
Ladies and gentlemen, I present you the next generation of cheating now available on console, and has been for a while but lately its been becoming more popular and more of a trend, consoles are no longer a safe space to play your games legit anymore pic.twitter.com/iEQzPVFf1h
— Anti-Cheat Police Department 🕵️ (@AntiCheatPD) July 5, 2021
Those who discovered the system say the player captures video feed from the console and transmits it to a PC. On the PC, a cheat program with machine-learning algorithms processes the signal and analyzes the scene. Based on this data it sends controller commands that automatically aim at an opponent and fire.
According to enthusiasts, game and console developers will find it difficult to detect such systems. Yet Sony and Microsoft are actively working on anti-cheat systems and patenting the tools.
Also on ForkLog:
- Chinese developers surpassed the United States in the international smart city competition.
- Israel deployed a swarm of AI-powered drones in real combat.
- Tencent employed facial recognition to track child gamers, playing at night.
- Russia to develop standards for artificial intelligence.
- Scientists developed AI that classifies thousands of galaxies in a second.
Subscribe to ForkLog's AI news on Telegram: ForkLog AI — all AI news!