What Is Ground Truth?
Ground Truth is the reference data, treated as correct, against which an AI model's predictions are trained and evaluated.
Ground Truth — the reference data, treated as correct, against which an AI model's predictions are trained and evaluated.
Ground truth is the "right answer" a model is measured against — for example, the verified labels in a dataset. Its quality is decisive: if the ground truth is biased, incomplete, or wrong, the model will faithfully learn those flaws, which is why scrutinising how ground truth was created is a key governance question.
Source: Machine-learning practice
Plain-language explanation
Ground truth is the "right answer" a model is measured against — for example, the verified labels in a dataset. Its quality is decisive: if the ground truth is biased, incomplete, or wrong, the model will faithfully learn those flaws, which is why scrutinising how ground truth was created is a key governance question.
Related terms
See where you stand on AI governance
Take the free 7-question maturity assessment and get a personalised action plan.
Free assessment — 3 minutes →