Transcript: Memory Contagion: Cross-Temporal Propagation of Evaluator Bias via Agent Memory

Alex: Welcome to another episode of ResearchPod. Today, we're looking at a problem researchers are calling "Memory Contagion" — the idea that biased information can quietly spread through an AI system over time, carried forward by the system's own stored experiences.

Sam: So the paper is asking whether an AI can essentially pick up bad habits from its own past?

Alex: That's the core question. When an AI learns from feedback that has some kind of slant or bias baked in, it stores those experiences. Then, when a future version of that system goes looking for guidance on how to behave, it pulls from that same pool of stored memories — and the bias comes along for the ride.

Sam: That sounds like a serious design problem. If the system is supposed to be helpful and fair, but its memory is full of skewed examples, does it just start copying those patterns without realising?

Alex: That's exactly what the researchers found. And here's the part that makes it tricky: even if the system is perfectly organised — even if it files and retrieves memories without any technical errors — the bias still spreads. Because the problem isn't how the memories are stored. It's what's in them.

Sam: So it's not a filing error. It's a content problem.

Alex: Right. Think of it like a student who only ever studies from textbooks written by someone with a very particular, narrow point of view. The student might be an excellent reader with a perfect memory. But if every source they've ever encountered carries the same slant, they'll absorb that slant too — not because they're careless, but because the material itself is skewed. In this case, the "textbooks" are the stored experiences of previous AI agents.

Sam: And the AI uses those experiences as a kind of template for how to behave next time?

Alex: Exactly. The researchers call this "Memory Contagion." The system retrieves past examples to help it decide what to do, and if those examples are biased, the bias gets baked into its future decisions.

Sam: How did they actually measure whether the bias was spreading? You can't just ask the AI if it's picked up bad habits.

Alex: They used a mathematical tool that measures the distance between two sets of behaviours — think of it like measuring how much effort it would take to reshape one pile of sand into a different shape. If the AI's responses, taken as a whole, have drifted toward a particular pattern, that tool can detect it, even if no single answer looks obviously wrong.

Sam: So they're looking at the overall shape of the AI's behaviour, not just individual answers.

Precisely. That's important, because a single biased response might look like a one-off. But when you look at the whole distribution — the full range of how the system tends to respond — a shift becomes visible.

Memory Contagion: Cross-Temporal Propagation of Evaluator Bias via Agent Memory

Zewen Liu

Jun 22, 2026·6:23·arxiv·Original Paper

Memory ContagionEvaluator BiasAgent Memory ConsolidationCross-Temporal PropagationΓ temporalContent Contamination

Listen

6 min

Briefing

Research Question

This paper investigates whether evaluator bias—systematic preferences in reward models or human feedback—can persist and propagate across time in autonomous LLM agents through their memory systems. The author introduces the concept of "Memory Contagion," defined as the cross-temporal propagation of bias from past experiences to future agent behavior via stored memory.

Approach

The study formalizes Memory Contagion using a metric called Gamma_temporal, which measures the behavioral distance between agents retrieving from biased versus clean memory stores. The author conducts a four-phase experimental pipeline using two types of bias (length preference and authority bias) and compares two consolidation strategies: oracle consolidation (perfect, lossless merging) and LLM-based consolidation (summarization). The research further decomposes the contagion into content-based and retrieval-based components and performs a dose-response analysis to determine if there is a safe threshold for memory contamination.

Main Findings

The study reveals that Memory Contagion is a significant vulnerability in current agent designs. Key findings include:

Bias persists despite perfect consolidation: Even with oracle consolidation, length bias propagates to future agents, proving that biased input is a sufficient cause for contagion.
Bias-type and model dependency: Length bias propagates robustly in older models (DeepSeek V4-Chat), whereas authority bias failed to propagate in 15 controlled multi-seed experiments. Newer models like DeepSeek V4-Pro and Claude Sonnet 4.6 appear immune to length-based contagion.
No safe threshold: Contagion is detectable at contamination rates as low as 20% (p=0.2), suggesting that even small amounts of biased data can influence future agent behavior.
Mechanism: Contagion is driven by both the content of stored memories and the retrieval mechanism, with content-based contamination playing a slightly larger role.

Why It Matters

As autonomous agents are increasingly deployed in high-stakes environments, their reliance on long-term memory makes them susceptible to "historical" biases. This work demonstrates that simply improving memory consolidation is insufficient to prevent bias propagation if the input experiences are already tainted. It provides a formal framework for developers to measure and potentially mitigate cross-temporal bias in agentic systems.

Key Terms and Definitions

Memory Contagion — The cross-temporal propagation of evaluator bias through an agent's memory store to future agents.
Oracle Consolidation — A baseline memory merging process that preserves all information without loss, used to isolate the effects of biased input from consolidation artifacts.
Gamma_temporal — A quantitative metric using Wasserstein distance to measure the behavioral distance between agents retrieving from biased versus clean memory stores.
Length Bias — A systematic preference for longer responses, operationalized as a bias function favoring higher token counts.
Authority Bias — A systematic preference for responses containing authoritative markers, such as citations or institutional references.

Abstract

Large Language Model (LLM) agents increasingly rely on memory systems to maintain long-term coherence. Recent work shows that agent memories degrade during continuous consolidation. However, existing research assumes memories are derived from unbiased experiences. In this work, we identify and formalize a novel phenomenon: Memory Contagion -- the cross-temporal propagation of evaluator bias through agent memory. We show that when agents are trained or guided by biased evaluators, their experiences become biased; when these trajectories are stored and consolidated into memory, the bias propagates to future agents retrieving from the same memory store, even when consolidation is perfect (oracle). Across two bias types (length preference, authority bias) and four experimental phases, we demonstrate: (1) Memory Contagion occurs for length bias even with perfect consolidation on older models (Gamma_A = 13.18, DeepSeek V4-Chat), while newer models (V4-Pro, Claude) are immune, proving both that biased input is a sufficient cause and that contagion is model-generation-dependent; (2) authority bias fails to propagate in all 15 controlled multi-seed experiments (Gamma_A = 0.00), revealing that not all evaluator biases can cross temporal boundaries through current memory architectures; (3) No observed safe threshold: length bias propagation is detected at contamination rates as low as p=0.2. Our findings expose a critical but contingent vulnerability in current agent memory designs and provide formal tools for measuring cross-temporal bias propagation.

Memory Contagion: Cross-Temporal Propagation of Evaluator Bias via Agent Memory | Zewen Liu | ResearchPod

Memory Contagion: Cross-Temporal Propagation of Evaluator Bias via Agent Memory: Research Paper Podcast