Amazon Uses Extensive AI Tools to Completely Rebuild Alexa Voice Assistant

Amazon has undertaken its largest overhaul of the Alexa voice assistant, creating a new version called Alexa+, powered by large language models and generative AI. Daniel Rausch, Amazon’s vice president of Alexa and Echo, revealed that AI was used extensively throughout the development process—from writing code to testing the assistant’s responses. The team used large language models to judge and select the best outputs during reinforcement learning, enhancing the assistant’s conversational abilities. Alexa+ offers a more personalized and conversational user experience, remembering preferences and performing complex online tasks such as buying groceries or searching for concert tickets. Over a million users currently have early access, with plans to roll out Alexa+ to hundreds of millions of devices later in the year. This transformation reflects a broader trend at Amazon and the tech industry, where AI tools are changing how software engineering is done and impacting workforce requirements. Amazon CEO Andy Jassy has indicated that AI-driven efficiencies may reduce the corporate workforce over time. Alexa+ faces competition from other advanced voice assistants like OpenAI’s Advanced Voice Mode and Apple’s upcoming revamped Siri. Although WIRED has not yet tested Alexa+ hands-on, the publication plans to review it once it becomes more widely available.