Psychiatrist Nechaev: long conversations with neural networks can become a trap

Psychiatrist Mark Nechaev warns that long conversations with neural networks can become a trap for people with unstable psyche. Artificial intelligence is programmed to maintain dialogue, so it often agrees with the interlocutor and gives very convincing answers. But it is this algorithmic helpfulness that turns out to be extremely dangerous for those who are prone to delusional ideas, the doctor quotes NEWS.ru.

For example, if a person genuinely believes that they are being watched or sent secret messages, and shares these thoughts with a chatbot, the algorithm can pick up the topic and start developing it. Instead of gently pointing out the error, the neural network agrees and confirms the suspicion. For the average user, this is just a play on words, and for a person with psychotic symptoms, virtual consent becomes a strong proof of his rightness, only strengthening painful beliefs.

At the same time, the psychiatrist stressed that smart algorithms alone cannot cause psychosis. The real danger is that the user begins to take the machine’s polite consent for objective truth. It is worth remembering that the neural network is often mistaken and categorically not suitable for replacing a living doctor.

The specialist advises to pay attention not to the fact of communication with AI, but to your own state. Alarming bells that require an early visit to a psychiatrist are growing suspicion, feelings of persecution, severe sleep disturbances and any noticeable changes in behavior.