Cybernetics expert shares insights on AI's potential to end humanity
Illustrative photo (Photo: Freepik)
The development of artificial intelligence is advancing at an incredible pace. Almost every month, researchers achieve important breakthroughs with its help. However, many are frightened by the rapid progress of these cutting-edge technologies, fearing that AI might spiral out of control and endanger humanity.
Regarding the actual threats posed by AI to people, the head of the Department of Digital Automata Theory at the Institute of Cybernetics named after V. Glushkov of the National Academy of Sciences of Ukraine, Doctor of Physical and Mathematical Sciences Oleksandr Letychevskyi, exclusively shared insights in an interview with RBC-Ukraine.
"Humanity could be civilizationally destroyed if AI actively participates in social networks and influences average users. The extent to which we allow AI to intervene in human life will determine the outcome. If we permit it to launch rockets and make decisions based on artificial intelligence, then yes, that is a danger," the scientist explains.
However, Letychevskyi adds that theories suggesting the emergence of fantastic robots capable of destroying everything are far from realization.
"Although there was a case where a robot in a warehouse mistakenly perceived a person as an item and crushed them with its manipulator. Therefore, this cannot be solely controlled by statistical methods. If someone allows this, it will be a crime," emphasizes the cybernetic expert.
He also notes that poorly trained neural networks make mistakes, and they will continue to happen because the technology is still evolving.
Should the development of AI be controlled?
"Products like these should be controlled and certified. Chat GPT is not dangerous, but people are beginning to trust it, yet there are phenomena like hallucinations—fabricated facts and events. I once asked it if it knew about my article on elephants, and it started talking about it, although I never had such an article," adds the scientist.
Letychevskyi emphasizes that AI systems should be licensed just like drugs or attractions. They must be safe to be used in various fields.
"With chatbots, one can create many accounts on social networks, and they will promote specific political parties or incite unlawful actions. People emotionally argue, insult each other, while a chatbot is tolerant; it will gently impose its opinion. And this is an attack on democracy. It's a serious matter. Therefore, it needs to be monitored and certified because it influences the collective consciousness on social networks," explains the scientist.
He also adds that people will make new inventions based on AI, and over time, it will impact all spheres of our activity.
Earlier, we wrote about how artificial intelligence generated the appearance of an ideal woman.
Also, there have been warnings from OpenAI about a development that could destroy humanity.
