Today, Amazon unveiled a fresh suite of generative AI tools tailored for sellers, with the aim of streamlining the creation of product listings. These cutting-edge capabilities, according to the retail titan, are intended to assist sellers in crafting compelling product descriptions, titles, and listing particulars.
Additionally, sellers will have the capacity to augment their existing product descriptions using AI, eliminating the need to commence from scratch.
These AI tools were constructed using large language models (LLMs) that underwent extensive training on copious amounts of data. Although Amazon doesn’t explicitly state it, it appears that the retailer likely drew from its own troves of listing data to educate its machine learning models. Previously, Amazon had harnessed machine learning and deep learning methodologies to extract and enhance product information. However, these novel generative AI capabilities represent a significant advancement in this technology.
Robert Tekiela, Vice President of Amazon Selection and Catalog Systems, elucidated, “With our new generative AI models, we can deduce, enhance, and enrich product knowledge on an unparalleled scale, with remarkable improvements in quality, performance, and efficiency. Our models acquire the ability to deduce product information through diverse sources of data, latent knowledge, and logical reasoning. For instance, they can infer that a table is round if the specifications include a diameter or deduce the collar style of a shirt from its image.”
Amazon asserts that its generative AI tools will enable sellers to save time and empower customers to discover more comprehensive product information. Nevertheless, there are certain apprehensions surrounding the deployment of generative AI models at such a magnitude, given their potential to “hallucinate”—that is, generate false information not grounded in actual data.
These tools could potentially harbor other inaccuracies that may go unnoticed if not reviewed by a human. If the tools end up generating incorrect product listings and descriptions, Amazon could potentially bear legal responsibility, especially if it fails to disclose that AI was used in creating the listings.
Previously, The Information had reported Amazon’s pilot of generative AI tools for content, mentioning that the tool advises sellers to double-check content for compliance with Amazon’s listing guidelines. The report indicated that Amazon declined to provide details about the LLMs employed in the new tool.
Amazon is not the sole retailer turning to generative AI to simplify the process of crafting product listings. eBay recently announced the launch of a generative AI tool capable of generating product listings from photos. Earlier in the summer, Shopify introduced its own ChatGPT-like assistant for e-commerce merchants, designed to comprehend and interpret queries or prompts related to business decision-making and create content such as blog posts, campaign ideas, and customer emails, among other tasks.
