| Category | Technique | Description | Case Study | Defensive Measure |
| Injection Attacks | Insertion of malicious data | Deliberately adding harmful data to the training set, leading to erroneous model training. | Financial Forecasting Misguidance | Robust Data Validation, Anomaly Detection |
| Modification Attacks | Alteration of existing data | Subtly changing critical data points within the dataset, thereby corrupting the model’s output. | Manipulation of Facial Recognition | Data Provenance, Statistical Data Integrity Checks |