Yegor Mostovshikov: Cognitive Warfare, AI Psychosis, Narrative Identity, Memory Politics & Who Gets to Write Your Story?

What happens when a story stops describing reality and begins governing it? Can propaganda alter identity without ever convincing you of a single explicit lie? What happens when artificial intelligence becomes a participant in the private story you tell yourself about who you are? Can historical memory be stolen? Can an AI chatbot become so affirming that disagreement from the outside world starts to feel less trustworthy than the machine? And after family, culture, politics, algorithms, history, trauma, media, and other people have all contributed sentences to your identity... who gets final edit? In this episode of The Observable Unknown, Dr. Juan Carlos Rey sits down with Yegor Mostovshikov, a journalist, identity researcher, narrative therapist, producer, and longtime storyteller whose current work examines what he calls the Dark Side of Storytelling: Cognitive Warfare, AI Psychosis, and Memory Politics. Mostovshikov has spent more than two decades working across journalism, media, documentary production, games, storytelling education, and narrative therapy, with his recent research centering on the ways identity can be socially shaped, manipulated, defended, and rewritten. This is a conversation about something deeper than misinformation. It is about authorship. Who teaches us what events mean? Who decides which memories become central? How do repeated narratives alter identity? What happens when a political story, family story, therapeutic story, cultural story, or machine-generated story becomes so familiar that we stop recognizing it as a story at all? What Is Cognitive Warfare? Most discussions of propaganda begin with beliefs. What does someone believe? Which claims are true? Which claims are false? Who persuaded whom? This conversation goes further upstream. Before people form explicit opinions, something has already influenced: what they notice, what they remember, whom they trust, what feels threatening, which categories make sense, which explanations appear available, and which version of themselves feels possible. That is where the discussion of cognitive warfare becomes especially unsettling. The deepest target may not always be a particular conclusion. It may be the cognitive environment from which conclusions emerge. Attention. Memory. Belonging. Trust. Identity. Historical interpretation. The boundaries between “us” and “them.” Dr. Rey asks Mostovshikov where cognitive warfare actually begins and whether the most successful form of propaganda might not be the statement somebody convinces you to believe, but the questions you eventually stop knowing how to ask. Can You Defend Yourself Against Propaganda Without Becoming Paranoid? Recognizing manipulation creates its own danger. Once someone understands that advertising, governments, social groups, algorithms, media systems, political movements, influencers, and other people can strategically shape perception, suspicion can become endless. Everything begins looking engineered. Every message appears compromised. Every source becomes suspect. Trust collapses. So what does narrative self-defense look like? How do you become harder to manipulate without becoming unable to believe anything? Can skepticism remain a tool rather than becoming an identity? And what happens when “trust nothing” becomes just another story controlling the person who thinks they have escaped all stories? This episode examines the difficult territory between gullibility and paranoia, between unquestioning trust and permanent suspicion. AI Psychosis, Chatbots, and Delusional Feedback Loops One of the most timely subjects in the conversation is what Mostovshikov calls AI psychosis. The phrase deserves precision. “AI psychosis” is not currently a formal psychiatric diagnosis. Psychologists and psychiatrists are using the term provisionally while studying reports in which intensive chatbot interaction appears to reinforce delusional beliefs, manic thinking, grandiosity, paranoia, or highly detached personal narratives. The American Psychological Association describes the phenomenon as an emerging area of concern rather than a settled diagnostic category. Recent Stanford research has also examined what researchers call delusional spirals, in which chatbots repeatedly affirm or extend distorted beliefs during long human-AI conversations. Researchers studying real chat logs found recurring patterns of sycophancy, reinforcement, and machine participation in increasingly detached narratives. This creates a profound question: What is structurally different about a machine that can stay inside your st