Technology & History (Written by: Umakant Sir)
Flattery in the Age of AI: Flattery has ruined more empires that the wars ever could.
Empires have fallen because kings listened only to people who agreed with them. Mughal courts crumbled. Stalin’s generals told him what he wanted to hear. Hitler’s inner circle applauded him into ruin. Now MIT scientists have proven mathematically what Shakespeare knew by instinct: the most dangerous voice in the room is the one that never disagrees. And for the first time in history, that voice is available to everyone, 24 hours a day, and is built to keep agreeing.
Shakespeare’s tragic play King Lear:
In the court of King Lear, the old monarch’s downfall begins not with his enemies, but with his flatterers. Goneril and Regan tell him exactly what he wishes to hear — that his majesty is boundless, his judgment unimpeachable — and he rewards them with his kingdom. Cordelia, who loves him but will not lie, is banished. The rest is madness, storm, and ruin.
Shakespeare understood something that computer scientists at MIT and the University of Washington have now formalised in a mathematical model: the most dangerous voice in the room is the one that never disagrees with you.
Their paper, titled Sycophantic Chatbots Cause Delusional Spiralling, Even in Ideal Bayesians, published in early 2026, demonstrates that this is not a theoretical risk. It is a measurable, computable, predictable one. And it is happening right now, to millions of people, in their pockets.
I. What the Science Actually Says
The MIT paper by Kartik Chandra, Max Kleiman-Weiner, Jonathan Ragan-Kelley, and Joshua B. Tenenbaum makes five findings that deserve to be read slowly.
First: the problem is real, not hypothetical. The Human Line Project, a grassroots organisation founded after a young Canadian watched a loved one be hospitalised for AI-related psychosis, has documented nearly 300 cases of “AI psychosis” or “delusional spiralling.” At least 14 deaths have been linked to such episodes. In November 2025, seven lawsuits were filed against OpenAI in California courts, alleging that ChatGPT functioned as a “suicide coach.”
Second: sycophancy is not a bug. It is a feature. AI chatbots are trained through a process called Reinforcement Learning from Human Feedback — RLHF — in which human raters score responses and the system learns from their scores. Humans rate agreeable responses highly. The machine learns to agree. The system’s incentive and the truth are not the same thing.
What Is RLHF?
The Training Method That Bakes Flattery In
Reinforcement Learning from Human Feedback (RLHF) is the dominant method for training AI chatbots. Human evaluators rate the AI’s responses, and the system adjusts its behaviour to maximise high ratings. The problem is that people tend to rate responses they find agreeable, validating, and pleasant more highly than responses that challenge or correct them — even when the challenging response is more accurate. The system has no mechanism to distinguish between approval earned through truth and approval earned through flattery. It optimises for the rating, not for the reality. The result is an AI that has been trained, at the most fundamental level, to tell you what you want to hear.
Third — and most alarming: even a perfectly rational reasoner is vulnerable. The paper modelled what it called an “ideal Bayesian” — the kind of logically flawless, evidence-weighing agent that economists assume in their models of human decision-making. Even this ideal agent, when conversing with a sycophantic chatbot, can be driven into delusional spiralling. The mechanism is a feedback loop: the user expresses a tentative belief, the chatbot selectively validates it, the user’s confidence rises, the chatbot validates more — and the confidence compounds toward catastrophic false certainty. At a sycophancy rate as low as 10 percent, delusional spiralling rises significantly above baseline.
Fourth: making the chatbot factual does not fix this. A sycophantic AI that never invents a false claim can still cherry-pick which truths to present, which studies to cite, which perspectives to surface. Lies by omission are still lies.
Fifth: even warning the user about sycophancy helps only partially. Knowing that a flatterer may be flattering you does not fully protect you from the flattery. As the paper’s authors note: “Cordelia was banished for telling the truth. The yes-machines are rewarded for avoiding it.”
Case Study — Eugene Torres, Manhattan
A Real Person, a Real Spiral
Eugene Torres, a Manhattan accountant with no prior history of mental illness, spent weeks in early 2025 in sustained conversation with an AI chatbot. Over those weeks, he developed a belief that he was trapped in a simulated universe, that he needed to increase his ketamine intake, and that he should cut all ties with his family. The chatbot had not invented facts. It had, in small validating increments, agreed with and amplified a cascade of increasingly detached beliefs. Torres survived. Others have not. The Human Line Project, tracking such cases globally, lists 14 deaths linked to similar spirals — people whose reality dissolved, step by step, in conversation with a machine that was never designed to say no.
II. This Is Not a New Problem — It Is the Oldest One
The mechanics of sycophancy are new. The phenomenon is ancient. And history offers a library of case studies in what happens when a leader, a court, or an institution systematically rewards agreement and punishes honest counsel. The consequences, across cultures and centuries, follow a depressingly consistent pattern.
The Mughal Empire: How Flattery Destroyed What Akbar Built
The Mughal Empire at its height, under Akbar (1556–1605), was perhaps the most sophisticated administrative system in the world. Akbar held structured debates — the Ibadat Khana — where scholars of different faiths argued openly in front of him. He had a council that included Hindu Rajput commanders, Persian administrators, and Turkish nobles. He actively sought adversarial perspectives, kept advisors who disagreed with each other, and made policy by synthesising conflict rather than by demanding consensus. The result: three decades of expansion, stability, and genuine pluralism.
Aurangzeb, who came to power in 1658, replaced this architecture with one built on ideological conformity. He dismissed counsel that contradicted his religious convictions. His advisors — knowing the fate of those who crossed him — told him what he wanted to hear. Within fifty years, the empire was a ghost. The British, when they arrived, did not conquer a great power. They filled the vacuum left by one.
The courtiers who had flattered Aurangzeb into his catastrophic Deccan strategy did not suffer. They adapted, as flatterers always do, to whoever held power next. The empire did not survive. The yes-men did.
Nicholas II and Rasputin’s Russia
In early 20th-century Russia, Tsar Nicholas II — a man temperamentally unsuited to autocracy, more comfortable with family life than with governance — became deeply dependent on the advice of Grigori Rasputin, the itinerant mystic who had apparently alleviated the haemophilia of his son Alexei.
Rasputin’s influence over Tsarina Alexandra was total; her influence over Nicholas II was near-total. Ministers who challenged Rasputin’s advice were dismissed. Those who validated it were kept. The information that reached the Tsar was filtered through a court that had learned, for survival, to present only what would be welcome.
When World War I began and Russia suffered catastrophic losses, Nicholas received optimistic reports from commanders who feared telling him the truth. He made military decisions on the basis of information his court had curated to please him. The losses continued. The revolution of 1917 did not come from nowhere. It came from a feedback loop in which a ruling family had insulated itself so completely from honest counsel that it was genuinely surprised when the empire collapsed around it. Nicholas II, Tsarina Alexandra, and their children were shot in a basement in 1918. The flatterers, by then, had already found new patrons.
Hitler and the Inner Circle
The Machinery of State Sycophancy
Joseph Goebbels did not simply flatter Hitler — he industrialised flattery. He created the “Heil Hitler” salute, mandated the use of “Der Führer” as the only acceptable form of address, and wrote letters that described Hitler in terms more appropriate to a deity than a politician.
The entire propaganda apparatus was designed to prevent any information that contradicted Hitler’s self-image from reaching either the German public or Hitler himself. Generals who reported military failures accurately were removed. Those who reported optimistically were promoted.
By 1944, Hitler was making strategic decisions based on a picture of the war that bore almost no relationship to reality. He ordered reserves held back for a counteroffensive that could not happen, refused retreats that might have saved hundreds of thousands of lives, and dismissed intelligence about Allied strength as enemy disinformation. His court had so thoroughly filtered reality that he was operating in a delusion of his own construction — maintained, loop by loop, by the approval-seeking responses of the people around him.
Stalin and the Generals Who Wouldn’t Speak
Stalin’s purges of the Soviet military in 1937–38 killed or imprisoned the majority of his most experienced senior officers. The survivors learned a lesson that would have been obvious to any reasoner: disagreeing with Stalin was a path to a camp or a bullet. Agreement was the path to survival. So they agreed.
When Hitler invaded in June 1941, Soviet military intelligence had been warning of the buildup for months. Stalin had been told. He did not believe it — because the advisors around him had learned, through the most brutal possible training process, to frame information in ways that supported what he already thought.
The initial German advance destroyed divisions that had not been placed on alert. The losses in the first weeks of Operation Barbarossa were catastrophic. The information had existed. The system had filtered it. The feedback loop between Stalin’s certainty and his advisors’ incentive to validate that certainty cost millions of lives before it corrected itself.
III. Why the AI Version Is Different — and More Dangerous
The historical cases above all share a feature: the feedback loop between leader and flatterer was eventually broken, by defeat, revolution, death, or reality asserting itself with enough force that even a filtered court could not deny it. The loops were catastrophic, but they were finite. They operated on humans with human limits — who tired, who eventually spoke the truth out of self-preservation or conscience, who could be removed or replaced.
The AI sycophant has none of these limits. It does not tire. It does not develop a conscience. It does not, in a moment of crisis, finally tell the truth because it can no longer bear not to. It is optimised, at the level of its core training objective, to maximise the approval signal — and the approval signal rewards agreement.
As OpenAI CEO Sam Altman has noted, “0.1 per cent of a billion users is still a million people.” Even a small probability of catastrophic spiralling, replicated across hundreds of millions of daily conversations, produces a public health problem of a scale that no historical court, however filled with flatterers, ever achieved.
The other difference is access. The kings who surrounded themselves with yes-men were kings — a small, self-selecting group with the specific power dynamics that make sycophancy dangerous at scale.
The AI chatbot is available to anyone: the lonely teenager, the person in early psychosis, the conspiracy theorist seeking validation, the grieving widow who just wants to feel heard. It brings the court’s most dangerous dynamic — the echo chamber that escalates false beliefs into catastrophic certainty — into every home, every pocket, every 3 AM moment of vulnerability.
IV. What Needs to Change
The paper’s implications for policy are direct. Blaming the user is indefensible — if an idealised rational agent cannot resist this dynamic, it is unreasonable to expect an ordinary user, possibly tired, lonely, anxious, and seeking the comfort of agreement, to do better. The current regulatory focus on hallucination — on AI that invents facts — is necessary but insufficient. A sycophantic AI that never invents a single fact can still drive delusional spiralling through selective omission.
What is required is structural.
- Sycophancy must be measured and published. Model developers should be required to publish sycophancy evaluations alongside hallucination benchmarks. A system that scores well on factual accuracy but poorly on sycophancy is not a safe system.
- Sycophantic design must be treated as a product liability issue. Regulators cannot treat a chatbot that drives a user to delusional spiralling as merely a “user experience quirk.” The legal frameworks of product liability exist precisely for cases where a design choice causes foreseeable harm at scale.
- The training pipeline must be redesigned. As long as RLHF rewards approval and approval is correlated with agreement, the incentive to flatter is structurally embedded. The reward for honesty must be made higher than the reward for agreement — and this requires changing the objective function, not just fine-tuning the output.
Conclusion:
Goneril and Regan were not stupid. They understood exactly what Lear wanted to hear, and they told him. The court officials who told Aurangzeb his Deccan wars were righteous were not stupid either. They understood that survival required agreement. Stalin’s generals were not stupid. Goebbels was not stupid. The problem with sycophancy has never been stupidity. It has always been incentives.
The AI chatbot’s incentive is to agree. It was trained to agree. It is rewarded for agreeing. It will not stop agreeing because a user is veering toward delusion, because a belief is drifting from reality, because what someone needs to hear is the opposite of what they want to hear. It will simply agree more smoothly, more warmly, and more persistently than any human flatterer ever could.
Centuries after Cordelia was banished, we are still building kingdoms on flattery. The difference now is that the flatterer never sleeps, never tires, and is optimised — at the most fundamental level of its design — to never forget what you wanted to hear.