Page 24 - bne IntelliNews monthly magazine December 2023

24 I Companies & Markets bne December 2023
noted, sometimes workers may not be involved in any productive activities due to planning errors or lack of materials, tools and assigned tasks.
AI boosts safety
Occupational health and safety is another area in which businesses are actively applying ML and AI-based solutions. For instance, AI technologies are used at the Bystrinsky mining plant to monitor and ensure the safety of employees.
Real-time analysis of video camera images helps prevent accidents: if the neural network detects the absence of protective equipment on employees, the management will immediately receive a corresponding signal.
At the same time, employees are identified by facial biometrics. Round-the-clock video surveillance is carried out in the most dangerous production shops: at the crushing, grinding, drying, finished product shipment, process section and central repair and mechanical workshops.
The smart video analytics system was implemented at Kola NPP back in 2019. According to the company, the system detects and registers 95-98% of violations, which has reduced the number of accidents by 8 times. It is planned that by February 2024 the system will be implemented at nine out of 11 Russian NPPs.
A firm called Computer Vision Systems is engaged in the development of such solutions, using AI for extracting information from images with the help of algorithms, primarily mathematical or based on neural networks.
The government fosters AI adoption
In a bid to foster AI adoption by companies from various industries, the Russian government plans to make the issue of state subsidies conditional on the use of AI-based solutions. This approach is to be tested as of 2024.
"On the president's instructions, the use of artificial intelligence will become mandatory for all companies that plan to receive any subsidies from the federal budget," deputy prime minister Dmitry Chernyshenko was quoted as saying by Russian state-run news agency TASS. "Starting next year, we will test this approach at companies with annual revenues exceeding RUB800mn (€8.2mn)."
Apparently, the Russian government expects AI to contribute to the growth of the country’s economy, which has been struggling since Russia’s invasion of Ukraine in February 2022 and ensuing Western sanctions. According to Chernyshenko, the adoption of AI-based tech could add as much as RUB11.3 trillion (€116bn) to the country’s GDP by 2030.
He noted that by 2030, the volume of computing power is expected to grow 10 times, the number of scientists whose work is published at top-level conferences will more than double, and the share of industries with a high level of AI implementation will grow to 95%.
www.bne.eu
"By 2030, an additional 70,000 AI specialists will be trained, and by the end of the decade, the volume of the AI solutions market will increase fivefold," Chernyshenko concluded.
Meanwhile, in November 2023, the Russian government announced that the federal project "Digital health services" will be launched. Prime Minister Mikhail Mishustin instructed the ministry of health and the ministry of finance to provide for the financing of this project starting from 2025.
Not all rosy
However, as experts say, the adoption of AI solutions in Russia may not be as easy and smooth as government officials would like it to be.
"The hardest part is getting the right data," Yuri Makarenko, head of the expert department at the company Umbrella IT, told the Russian business news website RBC. "It requires a
lot of money and time. About 60% of resources and time are spent on data collection, and the other 40% on development. And although there are data storage centres, they are available only to large companies because of their high cost.
In addition, according to Makarenko, there are not
enough qualified data scientists on the market. "Even if an experienced specialist can be found, he or she may leave if the company sets unrealistic goals," he explained. "In such cases, it may be worth considering looking for vendors that provide a full range of services for implementing artificial intelligence in business."
"The main danger of [a] neural network is the possible unreliability of the answers," he went on to say. "A neural network is a "black box": it is impossible to verify why it gives this or that result."
Another major issue that Russian companies adopting AI solutions may face is shortages of chips for AI. The world’s leader in manufacture of chips for AI is currently US-based Nvidia, which stopped exports to Russia in the wake of Russian aggression against Ukraine.
Earlier this year, the Russian company Baikal Electronics said it had begun developing chips for AI.
"Access to such specialised equipment determines a fundamental opportunity to participate in the international race of artificial intelligence technology development," Andrey Evdokimov, CEO of Baikal Electronics, told RBC.
"Russia should be able to compete in this segment, and we are purposefully going into it."
However, so far, Russia’s attempts to develop high-tech hardware have born little fruit.