Recommendation poisoning is a cybersecurity threat targeting AI-driven recommendation systems, such as those used by streaming services or online retailers. Attackers intentionally inject biased, misleading, or malicious data into the system, aiming to manipulate the AI’s output. This can result in the promotion of harmful, irrelevant, or fraudulent content, undermining user trust and potentially causing financial or reputational damage. Recommendation poisoning can occur through fake user profiles, manipulated ratings, or fabricated interactions, making it challenging to detect. Defending against this threat requires robust data validation, anomaly detection, and continuous monitoring to ensure the integrity and reliability of AI recommendations.