A company is developing a new facial recognition system and is collecting a vast amount of images to train their machine learning model. To prevent compromises to the system's integrity, the security team must be vigilant against certain kinds of attacks that involve manipulating the training data set. Which type of attack specifically targets the integrity of the machine learning model by injecting maliciously modified data into the training set, with the intent to skew the model's future predictions or behavior?
Buffer overflow
Data poisoning attacks
Cross-site scripting
SQL injection