What risks does AI pose? Opinions from a cybersecurity expert and a psychologist

Artificial intelligence (AI) is rapidly transforming the world, entering everyday life, industry, and even the military sphere. However, behind its promising potential lie serious threats, including the loss of digital sovereignty and risks to personal privacy. Cybersecurity expert Kirill Malofeev shared his insights in an interview with Tsargrad.tv. Psychologist Ekaterina Orlova also warned about the dangers of using AI chatbots to discuss personal experiences. Here’s a look at the main threats posed by AI and how to minimize them.

Loss of Digital Sovereignty

According to Malofeev, the main threat of AI is the loss of control over information, which leads to a loss of digital sovereignty:

• Foreign Neural Networks: If a country does not regulate AI, it becomes controlled by foreign corporations or developers who shape data according to their interests. For example, algorithms may promote certain views or censor content, as seen on some Western platforms.
• Manipulation of Consciousness: AI that filters information (e.g., in search engines or social media) can influence public opinion, shaping narratives favorable to its creators.
• The Myth of Neutrality: “A neutral AI serving all of humanity doesn’t exist. Every model reflects the goals and biases of its developers.”

Malofeev quote: “If the state doesn't control what’s written in AI, someone else does. There will be no mythical neutral neural network.”

Example: In 2023, ChatGPT faced criticism for censoring certain topics based on OpenAI’s policies, sparking debates about free speech.

Solution: Malofeev emphasizes that countries are realizing the need to develop their own AI systems, social networks, messengers, and video platforms. For instance, Russia is advancing YandexGPT, VK Cloud, and RuTube to reduce dependence on Western platforms.

Privacy and Psychological Risks

Ekaterina Orlova, deputy director of the Institute of Clinical Psychology at Pirogov University, warned against using AI chatbots to discuss personal problems:

• Data Leaks: Deep fears, traumas, and personal experiences shared with chatbots can be accessed by hackers or corporations. In 2024, a data leak from the Grok (xAI) chatbot caused a scandal when user conversations were exposed online.
• Lack of Empathy: AI cannot analyze non-verbal cues such as facial expressions or tone of voice, making it unsuitable for real psychological support. Algorithms give templated responses that can worsen a person’s condition.
• Dependency: Frequent communication with chatbots can weaken real-world social skills and create dependency on AI.

Orlova quote: “By trusting your trauma to AI, you risk data leaks. An algorithm cannot replace a psychologist — it doesn’t see your soul.”

Example: In 2025, the Replika chatbot was criticized for collecting data on users’ mental states, which was later used for targeted advertising.

Other AI Risks

Beyond sovereignty and privacy, AI poses additional threats:

• Deintellectualization: According to Malofeev, dependence on AI weakens the ability to think critically and make independent decisions. A 2024 MIT study showed that 60% of students prefer AI solutions to critical thinking.
• Cyberattacks: AI can be used to create deepfakes, phishing messages, or hacking tools. In 2025, hackers used AI to attack banks by mimicking the voices of executives.
• Military Risks: AI in warfare (drones, targeting systems) increases the risk of autonomous errors, as seen in the 2023 failure of the U.S. Navy’s Aegis system, which mistakenly classified a civilian plane as a threat.
• Economic Threats: Automation through AI could eliminate up to 30% of jobs by 2030 (McKinsey forecast), especially in services and manufacturing.

How to Minimize Risks

Experts suggest the following measures:

• Government Regulation:
  • Develop national AI systems (e.g., Grok 3 in Russia via xAI API).
  • Create AI legislation, including data protection and algorithm transparency (similar to the EU AI Act).
• Digital Independence:
  • Build local platforms (social networks, messengers, cloud services).
  • Support domestic developers like Kaspersky or Yandex.
• Education:
  • Teach critical thinking and cybersecurity in schools.
  • Promote AI-related professions to reduce dependence on the U.S. and China.
• Privacy:
  • Use open-source or local AI models to minimize cloud-based data leaks.
  • Avoid sharing sensitive personal data with chatbots.

Example: In 2025, Finland introduced mandatory AI certification, requiring systems to be tested for bias and data security.

Prospects

• 2025–2027: Russia plans to invest ₽500 billion in AI development, including a national model based on Grok and YandexGPT.
• 2030: According to Gartner, 80% of countries will have their own AI platforms to safeguard sovereignty.
• Risks: Without global regulation, AI may intensify geopolitical conflicts — as seen in 2024 deepfake-driven cyberattacks.

Conclusion

AI is a powerful tool, but its risks — from loss of digital sovereignty to data leaks — require urgent attention. As Kirill Malofeev noted, without control, AI will serve its creators, not society. Ekaterina Orlova adds that chatbots cannot replace human interaction and may endanger personal privacy. Nations must develop their own AI systems, strengthen legal frameworks, and educate citizens in cybersecurity. Only then can AI become an ally rather than a threat to humanity.