In a groundbreaking move, Amazon has just introduced a cutting-edge set of generative AI tools tailored specifically for its sellers. These innovative tools promise to revolutionize the process of crafting product listings, offering sellers the opportunity to effortlessly generate captivating product descriptions, titles, and listing details. Notably, sellers can now augment their existing product descriptions using AI, eliminating the need to start from scratch.
These remarkable AI tools are underpinned by extensive training with large language models (LLMs) on copious amounts of data. While Amazon hasn’t explicitly detailed the sources of this data, it’s likely the retail giant harnessed its own extensive listing data to train these machine learning models. Building on previous machine learning and deep learning techniques employed for extracting and enriching product information, these new-generation AI capabilities represent a significant leap forward.
Robert Tekiela, Vice President of Amazon Selection and Catalog Systems, offered insights into the game-changing capabilities of these generative AI models, stating, “With our new generative AI models, we can infer, improve, and enrich product knowledge at an unprecedented scale and with dramatic improvement in quality, performance, and efficiency. Our models learn to infer product information through diverse sources of information, latent knowledge, and logical reasoning. For example, they can infer a table’s round shape if specifications list a diameter or deduce the collar style of a shirt from its image.”
Amazon anticipates that its generative AI tools will not only save sellers valuable time but also empower customers to access more comprehensive product information. However, there are valid concerns surrounding the use of generative AI models at such scale, particularly due to their potential to “hallucinate” – that is, create false information not grounded in real data.
There’s also the possibility of overlooked mistakes in the AI-generated content that only a human reviewer can catch. In the event that these tools produce inaccurate product listings and descriptions, Amazon could potentially face liability issues, especially if it doesn’t explicitly disclose that AI was involved in creating the listing.
Prior to this announcement, The Information reported that Amazon had been conducting pilot tests of generative AI tools for content creation, emphasizing that the tool prompts sellers to double-check content to ensure compliance with Amazon’s listing guidelines. The report also highlighted Amazon’s reluctance to provide specific details about the LLMs used for these tools.
It’s important to note that Amazon is not the only major retailer venturing into the realm of generative AI to streamline the process of crafting product listings. Just last week, eBay unveiled its own generative AI tool capable of generating product listings directly from photos. Earlier in the summer, Shopify introduced a ChatGPT-like assistant for its e-commerce merchants, designed to comprehend and respond to questions or prompts related to business decision-making. This assistant can create content such as blog posts, campaign ideas, and customer emails, showcasing the growing trend of harnessing AI for enhanced e-commerce experiences.
