{"id":18712,"date":"2026-11-05T07:00:45","date_gmt":"2026-11-05T06:00:45","guid":{"rendered":"https:\/\/www.moresophy.com\/?p=18712"},"modified":"2026-11-05T08:31:29","modified_gmt":"2026-11-05T07:31:29","slug":"when-agent-based-ai-plays-chinese-whispers","status":"publish","type":"post","link":"https:\/\/www.moresophy.com\/en\/blog\/when-agent-based-ai-plays-chinese-whispers","title":{"rendered":"When Agent-Based AI Plays “Chinese Whispers”"},"content":{"rendered":"\t\t
\n\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t
\n\t\t\t\t\t
\n\t\t
\n\t\t
\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t
\n\t\t\t\t\t\n\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t\tData-Driven Business<\/span>\n\t\t\t\t\t<\/span>\n\t\t\t\t\t<\/a>\n\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t
\n\t\t\t\t\t\n\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t\tGuest Column<\/span>\n\t\t\t\t\t<\/span>\n\t\t\t\t\t<\/a>\n\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t
\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t

Alexander Steiner<\/p>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t

\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t

|<\/p>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t

\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\tNovember 5, 2026\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t

|<\/p>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t

\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t
12 min. read<\/div>\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t

When Agent-Based AI Plays \"Chinese Whispers\"<\/h2>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t
\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t
\n\t\t
\n\t\t\t\t<\/div>\n\t\t
\n\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t
\n\t\t\t\t\t
\n\t\t
\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t
\n\t\t\t
<\/div>\n\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t
\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t
\n\t\t\t\t\t\n\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t\tResponsible AI<\/span>\n\t\t\t\t\t<\/span>\n\t\t\t\t\t<\/a>\n\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t
\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\tProf. Dr. Heiko Beier\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t

|<\/p>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t

\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\tNovember 5, 2026\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t

|<\/p>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t

\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t
12 min. read<\/div>\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t

When Agent-Based AI Plays \"Chinese Whispers\"<\/h2>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t
\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t
\n\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t
\n\t\t\t\t\t\n\t\t\t\t\t\t\n\t\t\t\t\t\t\n\t\t\t\t<\/path><\/svg>\t\t\t<\/span>\n\t\t\t\t\t\t\t\t\tBack to the blog overview<\/span>\n\t\t\t\t\t<\/span>\n\t\t\t\t\t<\/a>\n\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t
\n\t\t\t\t\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t<\/path><\/svg>\t\t\t\t\t\t\t<\/span>\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t<\/path><\/svg>\t\t\t\t\t\t\t<\/span>\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t<\/path><\/svg>\t\t\t\t\t\t\t<\/span>\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t\t\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t<\/path><\/svg>\t\t\t\t\t\t\t<\/span>\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t
\n\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t

What a Microsoft study on information loss in agent-based systems reveals\u2014and why companies shouldn\u2019t ignore it.<\/em><\/strong><\/p>

Everyone knows the game \u201cChinese Whispers\u201d: A message is passed from person to person, and even though everyone tries their best, what ends up being communicated is often different from what was originally said. In agent-based AI<\/strong>, this pattern repeats itself, only unnoticed: An agent takes information from another, builds on it, and passes it on. With each handoff, the information drifts a little further from the original.<\/p>

Yet agent-based AI systems<\/strong> actually promise the opposite. The idea is to automatically process information across entire process chains: one agent analyzes data, another makes decisions, and a third generates responses or triggers the next step in the process. For companies, this sounds like the logical next step toward greater efficiency and automation.<\/p>

A recent study by Microsoft Research has systematically examined this effect. Even the three most powerful language models distort, on average, a quarter of the original document\u2019s content after 20 consecutive processing steps on the same document; across all 19 models tested, the loss rises to about half. The damage isn\u2019t caused by spectacular malfunctions or obvious hallucinations, but rather occurs gradually and often goes unnoticed. The result remains plausible\u2014but it becomes increasingly inaccurate.<\/p>

What does this mean for companies that are using AI agents<\/strong> precisely to speed up processes?<\/p>

The widespread assumption: More delegation means greater efficiency<\/h3>

Expectations for AI agents are high. Many companies treat them like digital colleagues: a task is described once, and the process then runs largely autonomously. Because modern language models often perform well on individual tasks, it\u2019s natural to assume that this quality will hold up across longer process chains involving multiple agents, in which each agent builds on the work of the previous one.<\/p>

Researchers at Microsoft Research therefore investigated whether information is reliably preserved when a language model processes the same document over many steps. Each step builds on the result of the previous one. Does the content gradually deviate from the original, similar to what happens with human participants in a game of \u201cChinese Whispers\u201d?<\/p>

In the classic game of \u201cChinese Whispers,\u201d people usually laugh at the end because the message is obviously distorted. In agent systems, however, such deviations often remain hidden because the results still appear consistent, logical, and professionally worded. That is precisely why they are rarely questioned.<\/p>

What the study shows<\/h3>

The study \u201cLLMs Corrupt Your Documents When You Delegate\u201d<\/strong> by Philippe Laban, Tobias Schnabel, and Jennifer Neville (Microsoft Research) was published in April 2026 as a preprint on arXiv<\/a>. For the benchmark DELEGATE-52<\/code> The researchers created 310 work environments spanning 52 professional domains, ranging from accounting to program code to musical notation. Nineteen language models were tasked with repeatedly processing the same documents: first executing one instruction, then the opposite instruction, with the original output expected at the end. After 20 such iterations, the researchers compared the result with the original. Each processing step builds on the result of the previous one, similar to an agent chain in which each agent takes over the output of its predecessor. The results are alarming: In more than 80 percent of the tested combinations of model and subject area, the researchers observed what they call \u201ccatastrophic corruption\u201d\u2014a term they use when at least 20 percent of the original content is distorted. Even the best model tested performed reliably\u2014that is, with at least 98 percent agreement with the original\u2014in only 11 out of 52 subject areas.<\/p>

More important than the exact percentage is the fact that this loss usually remains invisible. The most powerful models, in particular, rarely delete content but instead modify it\u2014errors that are only noticeable upon close inspection. It\u2019s not obvious where the information has changed. Instead, the altered information travels through the entire chain and is used by every subsequent agent as the basis for the next decision. The end result is an answer that appears complete and sounds correct, even though the underlying knowledge base has already been eroded.<\/p>

Why Greater Autonomy Doesn\u2019t Solve the Problem<\/h3>

A key finding of the study contradicts a widespread assumption: Greater autonomy does not automatically lead to better results. When the researchers equipped the models with additional tools such as file access and code execution, the quality of the results did not improve\u2014on average, it actually deteriorated by another six percentage points.<\/p>

A second effect exacerbates the problem: The larger the documents being processed and the longer the chain of delegation, the faster the loss of information grows. These two factors do not simply add up; they reinforce each other. According to the study, the damage is not caused by many small inaccuracies, but by rare yet serious errors: Individual breakdowns account for about 80 percent of the total loss. The longer the chain, the more likely such a breakdown becomes, and each additional step builds upon the compromised result.<\/p>

This turns a technical detail into a business risk: The answers remain convincing, logically structured, and linguistically sound. The problem isn\u2019t their plausibility, but their origin. After all, plausibility isn\u2019t the same as evidence. From our perspective, this is one of the biggest challenges for companies that want to use AI agents<\/strong> productively. When no one can trace how a result was generated, it becomes increasingly difficult to determine whether an answer is actually correct or just sounds correct.<\/p>

What agent chains need: checkpoints and human-in-the-loop<\/h3>

Agents can orchestrate workflows and coordinate tasks, but they do not automatically guarantee that information remains accurate, complete, and consistent throughout the process. The Microsoft study notes that users must continue to closely monitor such systems. That does not mean automating less. However, an architecture is needed with checkpoints where information loss and errors are detected before they propagate through the chain of agents.<\/p>

Three building blocks are crucial here:<\/p>