09/22/2023
Data poisoning is a fraudulent practice that involves injecting false or biased information into the training data used by AI models.
The intention ??
To manipulate the learning process and lead to inaccurate results.
It impacts across industries and its consequences can be far-reaching, affecting decision-making and operations.
How to Prevent Data Poisoning?
Data Quality: Collect data from trusted and reliable sources
Data Monitoring: Regularly monitor and audit the training data
Data Cleaning: Remove any suspicious or potentially harmful data points