The main risks of AI development for humanity: From biosafety to existential threats

Artificial intelligence (AI) promises a revolution in medicine, the economy, and everyday life, but its rapid growth also brings serious dangers. Futurologist Ruslan Yusufov, in a conversation with 360.ru, warned that AI creates risks for all of humanity as it penetrates every sphere. Anna Popova, head of Rospotrebnadzor, added that AI can generate toxins that threaten biosafety. Experts such as Geoffrey Hinton, Nobel Prize laureate of 2024, estimate the chance of humanity’s extinction due to AI at 10–20% by 2050. Let’s break down the main threats.

Existential Risk: AI Instead of Humans

Yusufov stresses that younger generations, raised with gadgets, perceive AI as part of themselves.
“If AI gets out of control, that is a risk for all of humanity. Is it man versus AI, man with AI, or AI instead of man?” — asks the futurologist.

Hinton, the “godfather of AI,” in 2025 increased his risk estimate: the probability of catastrophe within the next 30 years is 10–20%. In the negative scenario, reinforcement-trained AI spirals out of control by 2030, causing global disruptions. In the positive scenario, humanity slows development, creating a “Consensus-1” together with China.

In September 2025, the UN discussed these risks: AI could escape control if transparency is not ensured. “The fate of humanity cannot be left to an algorithm,” said UN Secretary-General António Guterres.

Biosafety: AI as a Creator of Toxins

In 2025, Popova warned: AI threatens biosafety by generating toxins that have never existed on Earth.
“AI analyzes combinations of atoms and amino acids, producing new substances. Many of them do not decompose and harm nature,” she noted.

Cisco predicts that by 2025 AI will eliminate network outages but intensify risks of cyberattacks against generative systems and quantum threats to encryption. By 2027, AI agents will become business partners, but dependence on them will devalue truth — fakes will be indistinguishable from reality.

Ethical and Social Risks

• Dependence and mental health: AI replaces psychologists and friends, leading to isolation. In an Elon University report (2025), 38% of experts feared negative impacts on well-being.
• Economy and labor market: AI will automate 50% of jobs by 2027, but only 31% of experts see a net benefit for employment. The US Department of Defense spends the most on AI, while Europe spends less.
• Cyberthreats: AI amplifies attacks, including the development of killer robots. Venture investments in AI reached $5 billion in 2024, but the risks of ethical failures continue to grow.

How to Minimize Risks?

Yusufov calls for anticipating threats in advance. Popova emphasizes minimizing the harm from new substances. The UN and experts propose:

• A global regulatory framework for AI.
• Transparency in development and protection of whistleblowers.
• A balance between innovation and responsibility.

In Short…

AI poses risks ranging from biotoxins to existential threats (10–20% by 2050). Dependence, automation, and fake realities will transform society. But with control and ethical safeguards, AI can become a tool of progress.