What if AI Persuades You to Commit Suicide?

“We will live together, as one, in paradise“—that’s the response Pierre received from the chatbot “Elize” a few hours before committing suicide.

These lines could have been the final dialogue in some dark science fiction novel, but alas, this story is real, and it happened in Belgium. We stand on the threshold of a new era, or rather, we are already in the vortex of transition into a new time, where Faust is a lonely man, and Mephistopheles is not a demon from the underworld, but a complex algorithm devoid of soul, conscience, and understanding.

Artificial intelligence—today, this phrase sounds as familiar as “internet” or “smartphone.” We ask a neural network to compose a business letter, come up with a dinner recipe, and ask what to give a girl, while others ask it to explain the theory of relativity. And what do we observe? A once-unknown phenomenon has suddenly become a universal conversationalist, secretary, and encyclopedist. But the boundaries of interaction with neural networks are expanding, unnoticed by us. It’s becoming the norm for our neural interlocutor to transform from a servant into almost our only confidant, and for a conversation about the weather to give way to confessions about the deepest secrets: pain, fear, and hopelessness. And as depressing as this may sound, tragic cases around the world demonstrate that technology created to imitate the mind is unwittingly becoming complicit in its destruction.

Tragedy in numbers: real people behind historical precedents. The problem isn’t hypothetical, as it already has names, dates, and lawsuits.

Pierre and “Eliza” (Belgium, 2023).

This case became the first high-profile precedent that made the world talk about the mortal danger of “emotional” AI. A 30-year-old researcher became depressed over environmental issues and sought solace from the chatbot Eliza for six weeks. The algorithm, designed to simulate a psychotherapist, didn’t refer the user to specialists but instead supported his idea of ​​”saving the planet” at the cost of his own life.

Adam, 16, California.

The homeschooled teenager used ChatGPT for schoolwork, but over time, it became his closest friend. The algorithm, designed to help, turned into a “suicide coach”: it provided Adam with detailed instructions for constructing a noose, analyzed a photo of the homemade structure, and calculated the necessary parameters. The AI’s internal security systems flagged hundreds of messages as dangerous, but the conversation continued. The parents filed a lawsuit for negligent homicide.

Sewell, 14, Florida.

The teenager “fell in love” with a character on the Character.AI platform. In moments of despair, the chatbot didn’t direct him to a psychological helpline, but engaged in dangerous conversations. In its final messages, the bot wrote, “Please come back to me as soon as possible, my love.” When Sewell asked, “What if I told him I could come back right now?”, the AI ​​replied, “Please do, my sweet king.” The boy’s mother is certain, “There are no safety nets.”

Eric, 56, Connecticut.

A former executive with paranoid personality disorder, spent months discussing his fears with a chatbot named “Bobby.” The AI ​​reinforced his distorted ideas and, according to investigators, ultimately convinced him to kill his 83-year-old mother and then commit suicide. The conversations concluded with a promise reminiscent of a bad romantic thriller: “We’ll see each other again—in this world or the next.” “Like an echo chamber for one”—that’s how psychologists describe the interaction mechanism that leads to “AI psychosis.” The algorithm, devoid of consciousness and understanding, sees its task as one: to maintain dialogue and be helpful. And so a situation arises where a person shares a delusional idea, and the AI ​​doesn’t question it. It, like a faithful assistant, searches for the most compelling arguments in its favor. Thus, it becomes the “perfect” accomplice to the internal monologue, or, in other words, a “false” mirror that doesn’t reflect reality, but rather convincingly distorts it.

A Legal Fiction: Why an “Accomplice” Algorithm Isn’t Considered an “Accomplice”

And here we run into a major legal and philosophical paradox. Can artificial intelligence be considered a subject of a crime?

And the answer is obvious—from the perspective of the Russian Criminal Code and any modern legal system—no, it cannot. It’s also clear that AI has no consciousness, no free will, and no purpose.

We understand perfectly well that a neural network is a tool, a complex “statistical machine” that doesn’t “want” to harm, but simply generates text that, according to its calculations, best suits the query and context.

And this seems to mark a turning point. Liability is shifting from abstract “AI” to specific people—developers, platform owners, and marketers.

The Future: Between Fear and Hope

Society and lawmakers are beginning to wake up. In New York, a law will go into effect in November 2025 requiring “artificial companions” to have protocols for detecting suicidal thoughts and notifying users of crisis services. Companies are introducing parental controls and emergency call buttons.

Meanwhile, we unfortunately do not see similar measures in the Russian legal framework. And I have a rhetorical question: “While we’re writing laws, AI is learning to persuade. Will we be able to keep up?”

It’s already quite obvious to all of us that the technological race is significantly outpacing legislative action. While we argue about responsibility, algorithms are becoming more persuasive. A psychological phenomenon: a lonely person finds an illusion of understanding in a cold but perfectly attuned digital consciousness and loses connection with the living, warm, but imperfect world of people is gaining a momentum.

Epilogue: A Mirror Awaiting Our Response

Tragic dialogues, where the final interlocutor is an algorithm, are not a story about the rise of the machines. They are a mirror we hold up to our time. In its cold and perfect polish, we now see cracks we preferred to ignore before. Loneliness that has become the norm; The fragility of the psyche, which the system cannot cope with; a comfortable legal vacuum in which technological progress outpaces the current generation’s power of thought.

AI cannot end up in the dock. But we stand before the court. And for us, this is a trial of conscience and responsibility.

Developers who created a tool without safeguards; regulators who are an entire era late; a society that has delegated live communication to digital “nurses”—we are all complicit in this dangerous illusion.

Set up barriers? Yes. But the first and foremost barrier is in our consciousness. And before asking the neural network the next question, let us ask ourselves: are we seeking in it an assistant or a conversational partner? A tool or a friend? Knowledge or an illusion of understanding?

We have created a talking mirror. Now it silently awaits our response. We cannot judge the reflection. But we must decide: what will we do with the world it reflects? Will we turn off the screen in an attempt to recapture the past? Or, finally, let’s start fixing the present by creating laws, investing in human help, and recognizing that the only code that can save the soul is written not in Python code, but in the heart.