On August 11, 2025, President Kassym-Jomart Tokayev said that artificial intelligence (AI) should become the driving force behind Kazakhstan's development. In the next 5 years, the country plans to transform into a digital state that makes the most of the potential of AI. Kazinform has studied global examples of AI implementation.
Medicine and Healthcare
In Israel, artificial intelligence is actively used in healthcare for diagnostics, personalized treatment, emergency response, and drug development. For example, the startup Aidoc has deployed AI algorithms in the country’s largest hospitals to instantly detect strokes, hemorrhages, and fractures on CT scans.
Personalized medicine is also developing rapidly: a joint project between the Technion and Maccabi has reduced the number of incorrect antibiotic prescriptions for urinary tract infections by 35% and helped combat drug resistance.
AI is also being used in emergency services, such as United Hatzalah, where the system predicts the likely locations of emergency calls with up to 85% accuracy, reducing response times.
However, the use of medical data requires strict protection, as leaks can compromise patient privacy. The high cost of AI-based solutions and the need to integrate with existing systems make them difficult to implement in smaller clinics. There is also a risk that doctors will become overly dependent on AI, which could reduce their attention when making diagnoses.
To mitigate these risks, several countries have introduced standards for the certification of AI medical systems, mandatory algorithm validation procedures, and protocols for shared decision-making between doctors and AI. Open medical data is being created to improve the accuracy and objectivity of models. Training medical professionals in AI helps them correctly interpret results and perceive AI as a support tool rather than a replacement for clinical expertise.
Finland is actively introducing AI into education, combining a national strategy with innovations at school and university level. For example, the ViLLE platform, developed by the Institute for Learning Analytics in Turku, uses adaptive feedback methods. It analyzes students’ responses, strengths and weaknesses, and other indicators to determine where they need additional support and where they are ready for new tasks.
However, algorithms can be biased, such as incorrectly grading papers based on writing style or cultural context. Over-reliance on AI in teaching can weaken critical thinking and independent research skills. The introduction of AI can also increase educational inequalities between regions and social groups.
To mitigate these risks, many countries are creating ethical standards for the use of AI in education, including mandatory checks for transparency and fairness. Hybrid models are becoming increasingly common, in which AI supports teachers rather than replaces them. Teachers are trained to use AI tools so that they can effectively monitor the learning process and correct systemic errors.
Transport and LogisticsThe Netherlands is actively using AI to improve the efficiency and sustainability of its transport infrastructure. In the Port of Rotterdam, an AI system accurately predicts the arrival times of ships using data on past arrivals, vessel types, routes and speeds. This has reduced average waiting times by 20%, improving planning for terminals, shipping agents and ship operators.
In the US, AI is supporting logistics at the industrial service level. For example, Uber Freight is using machine learning algorithms to reduce the number of empty truck trips. Typically, around 35% of trucks travel empty, but this figure has been reduced by 10-15%.
However, forecasting systems are highly dependent on high-quality and up-to-date data, and many ports and logistics centers still operate with disparate or outdated information flows. Excessive automation of decisions poses operational risks when algorithms misinterpret unusual conditions, such as extreme weather conditions or geopolitical disruptions.
CybersecurityAI is increasingly being used to detect, prevent, and respond to cyberthreats in real time. In the US, Microsoft Security Copilot uses generative AI to help security analysts investigate incidents, correlate threat data, and develop remediation measures.
In the financial sector, HSBC uses AI models to monitor millions of transactions a day, identifying suspicious activity and blocking fraudulent payments in seconds.
However, AI-based security systems can themselves become targets for attack: attackers use adversarial attacks by feeding false data that causes the system to miss threats or generate false alarms. Models on
Scientists who rely on incomplete or biased data may fail to recognize new types of attacks, while overreliance on automated solutions can delay human intervention during complex incidents. However, attackers are also using artificial intelligence – to automate phishing attacks, create polymorphic malware, and search for network vulnerabilities en masse.
Energy and the environmentForecasting wind power generation is key for Denmark, where half of its electricity comes from renewable sources, and on some days wind accounts for up to 50% of consumption. AI-powered models significantly improve forecast accuracy, allowing the grid to better cope with erratic production.
In Australia, startup Neara is analyzing grid infrastructure, taking into account extreme weather conditions, helping to select optimal repair strategies and improve supply reliability.
While AI contributes to environmental and energy goals, its own development can harm the environment. Training generative models with billions of parameters requires enormous amounts of energy, increasing CO₂ emissions and straining electricity grids. Cooling servers also requires millions of liters of water.
To reduce the environmental footprint of AI, countries are developing greener data center solutions. Liquid and immersion cooling systems are increasingly being used, which reduce energy consumption by 50% and reduce water consumption. Data centers are also switching to renewable energy sources. In Brazil, for example, they are connected to a grid that is almost 90% hydroelectric.
AgricultureIn the Netherlands, greenhouses use computerized ripening systems to monitor plant health and automatically adjust watering, lighting, and temperature. In the US, farmers are using platforms like John Deere See & Spray, which uses AI to identify weeds and spray herbicides only in the right places, reducing chemical consumption by several times. In Australia, drones with AI analytics monitor the health of livestock and the condition of pastures.
But there are challenges here too. The high cost of equipment and the complexity of integration make it difficult to implement AI on small farms. Algorithms can make mistakes - for example, they can misdiagnose plant diseases or soil conditions, leading to losses. Reliance on cloud services and sensors increases vulnerability to cyberattacks that can disrupt entire farming operations.
To address these issues, governments and companies are launching training programs for farmers, providing subsidies for technology purchases, and creating artificial intelligence platforms adapted to local conditions. Hybrid systems are also being developed that can operate without a permanent internet connection, reducing the risk of damage and attacks.
The news was published on the basis of an information exchange agreement between Fakti.bg and Kazinform