Data Poisoning
Data poisoning is a malicious attack where bad or manipulated data is fed into an AI system to corrupt its training and make it produce incorrect or biased outputs.
AI systems learn from the data they process. Data poisoning is a type of cyberattack where an adversary intentionally introduces corrupted, inaccurate, or misleading data into an AI model's training dataset or its live operational data stream.
The goal of data poisoning can be to sabotage the AI's performance, introduce biases, or even trick the AI into making specific incorrect decisions. For small businesses, this can compromise the reliability of AI tools, lead to flawed business decisions, and damage customer trust or brand reputation.
An attacker could inject fake customer reviews with extremely negative sentiment into an AI sentiment analysis tool, causing it to misinterpret overall customer feelings about a product or service.