Skip to content
~/reflective_data

analytics-dictionary / term / ai

Model drift

Model drift is when an AI or machine learning system gets worse over time because the world or the data changes.

Written by Reflective Data experts · updated Oct 5, 2026

A model trained on past behaviour may perform poorly when customers, products or language change. Regular evaluation and monitoring catch it early.

For language model applications, changes in underlying models or content can have similar effects.

related

AI evaluations (evals)AI observabilityLarge language model (LLM)

// how we can help

We build company brains, agent harnesses and AI bots on top of data you can trust. AI services → · The data platform behind it

more in ai

reranking.md

Reranking

Reranking reorders retrieved results with a more precise model, so the best passages go to the…
token.md

Token

A token is a small piece of text, often part of a word, that language models…
semantic-search.md

Semantic search

Semantic search finds results by meaning rather than exact keywords.
embeddings.md

Embeddings

Embeddings are numerical representations of text or other data that capture meaning, so similar things are…
fine-tuning.md

Fine-tuning

Fine-tuning is training an existing model further on your own examples, to specialise its behaviour or…
ai-guardrails.md

AI guardrails

Guardrails are the rules and checks that keep an AI system within safe and intended behaviour.

← full analytics dictionary