Amazon’s Generative AI Investment Is Bigger Than You Think

While Google and Microsoft have often made headlines for their AI investments, Amazon appears to be a silent player stepping up its AI accelerator to capture business from the generative AI boom. The retail giant announced today it will begin using generative AI to help customers better surface product reviews and scale back on time spent rummaging through a series of individual reviews. Amazon will enable this by picking up common themes and summarizing them into short paragraphs on the product detail page. For example, a customer looking to understand whether a product is easy to use can surface reviews mentioning “ease of use” by tapping on that product attribute under the review highlights. The tool is currently accessible to a portion of mobile shoppers in the U.S., covering a wide range of products. According to the company's blog post, it plans to reach a broader audience and include more product categories in the upcoming months, based on customer feedback. The New York Times Updates Terms of Service to Prevent AI Scraping Its Content “Last year alone, 125 million customers contributed nearly 1.5 billion reviews and ratings to Amazon stores, making reviews an incredible resource for shoppers,” Vaughn Schermerhorn, director of community shopping at Amazon, told Adweek in a statement. “We want to make it even easier for customers to understand the common themes across reviews and ... we believe we have the technical means to address this long-standing customer need.” OpenAI’s launch of ChatGPT in November spurred tech giants like Microsoft and Google to act fast and invest in generative AI. Microsoft, who poured $13 billion into OpenAI, was quick to add LLMs into its own products such as ad-powered Bing earlier in February. The company also opens up revenue streams with a $30 monthly fee for this tool. Around the same time, Google too launched its own version of LLM—Bard—followed by an investment of $300 million in OpenAI rival Anthropic. The tech giant has also added generative AI to its search engine, but marketers are yet to see how this changes their search business. "Amazon is used to leading markets, and their relatively quiet position in this space indicates an understanding of the impact that generative AI can have on influencing customer buying choices," said Nicole Greene, senior director analyst in the Gartner Marketing practice. “Marketers need to pay attention to how Amazon consumers and their own customers are going to experience these efforts.” Amazon’s generative AI playground Amazon debuted in the AI race in April this year through its AI platform called Bedrock, which helps business scale their machine-learning models. This service extends to Amazon Web Services, which offers access to its first-party language models called Titan and other startups, such as Anthropic and Stability AI. In one such case, one Titan model can generate text for blog posts, emailers or other documents, per reports by CNBC. The other can help with search and personalization. AWS also invested $100 million in its Generative AI Innovation Center in June to connect AWS AI and machine learning experts with customers across the globe and guide them in building new generative AI products and services. Omnicom Group, among the first ad companies to work with this innovation center, will use AWS generative AI and machine learning to automate activities like developing creative briefs, media plans, ad creative, audience segmentation and performance measurement. However, Amazon’s AI ventures do not stop here. Reportedly, the retailer giant is also testing AI tools for sellers to craft titles and descriptions for product listings with minimal manual effort. To that, the tech giant is testing generative AI in its search capabilities for its ecommerce arm, per reports by Axios. The new feature—expected to roll out in the coming weeks--will be useful for aspects like improving the product recommendation engine, enabling personalized shopping experience, and enhancing customer data management. Elsewhere, the company is also designing its own chips Inferentia and Trainium that offers AWS customers an alternative to train their LLMs on Nvidia GPUs that are increasingly becoming expensive and difficult to procure, per CNBC reports. All this said the technology is still early in practical applications and direct consumer engagement. "It's one thing to build an internal model, it's another to allow it to communicate directly with customers," said Greene.