Overfitting
Definition: Overfitting happens when an AI model learns its training data too well and fails to handle new information correctly. It performs perfectly on data it already knows but struggles when faced with something different.
Example
If an AI is trained only on one law firm’s contracts, it might spot issues in those perfectly but miss important problems in contracts from other firms. It memorized patterns instead of learning how to adapt.
Why It Matters?
Overfitting makes AI systems unreliable in the real world. In law, that means an AI trained on limited case data could give poor advice or miss risks when analyzing new situations. Preventing overfitting ensures that AI tools make accurate, flexible, and trustworthy decisions.