Amazon relies on artificial intelligence (AI) to detect and eliminate fake customer reviews on its platform. When a customer submits a review, Amazon’s AI system analyzes it for known indicators of inauthenticity before publishing it online. While the majority of reviews pass this authenticity check and are posted immediately, some undergo further scrutiny. If the system identifies a review as fake, Amazon takes swift action, including blocking or removing the review, revoking the customer’s review permissions, blocking fraudulent accounts, and pursuing legal action if necessary.
Amazon’s machine learning models use proprietary data, considering factors such as seller advertising, customer reports of abuse, unusual behavioral patterns, review history, and more. Large language models (LLMs) and natural language processing techniques collaborate to identify anomalies in the data that suggest a review might be fake or influenced by incentives like gift cards or free products. Deep graph neural networks are employed to analyze complex relationships and behavioral patterns, helping detect and remove groups of bad actors.
Distinguishing between authentic and fake reviews can be challenging, as factors like rapid review accumulation due to advertising or genuine positive experiences may lead to misconceptions. Amazon also engages expert investigators when necessary to gather additional evidence for suspicious reviews. In 2022, Amazon blocked over 200 million suspected fake reviews globally. The company emphasizes its commitment to maintaining the authenticity and reliability of customer reviews on its platform.
