Data Poisoning
Data poisoning is when attackers inject corrupt or misleading data into ML training sets — teaching models the wrong patterns, biases or blind spots on purpose.
What is Data Poisoning?
Data poisoning is an attack technique where malicious actors inject corrupted or misleading data into machine learning training datasets. This contamination can compromise AI models, causing them to make incorrect predictions or fail to identify threats. Data poisoning is closely related to bot contamination, where automated traffic pollutes datasets used for analytics and machine learning.
Prosopo helps prevent data poisoning by blocking malicious bots and ensuring clean, authenticated traffic reaches your applications and data collection systems.