On Wednesday, September 20, the online commerce giant Amazon held its back-to-school show to tout its upcoming products and services. A busy program which gave pride of place to its voice assistant Alexa, which is boosted by generative AI.
A few weeks before Prime Days, the two big days of all-out promotions from the online commerce giant which will take place on October 10 and 11, Amazon has organized its big back-to-school show dedicated not to distribution but to its products and services. Among all the new features announced (read our article), the star of the day was none other than Alexa, the in-house voice assistant. Introduced in 2014 and embedded in more than half a billion devices around the world according to Amazon, Alexa should undergo a major facelift in the coming months to respond even better to the tens of millions of interactions per hour that she already assures. The goal: to make it even more natural, more involved and, obviously, more essential for its users. And Amazon seems to have found the recipe’s defining ingredient: generative artificial intelligence.
Amazon Alexa: a new friend to chat with
The generative AI that Amazon is integrating into Alexa is yielding spectacular results. It is possible to chat with the assistant in almost the same way as you would with another human being. We can ask him questions (this is already possible today) but also cut him off to add a clarification, continue the conversation without needing to call on him again, stammer, speak in a disjointed or jerky manner , etc. The assistant remains attentive and expresses itself more naturally, with ease and fluidity. Impressive.
To achieve this, the assistant requests all the components and data that it has at its disposal. If an Amazon Echo with a camera is used to call Alexa, she will know that you are speaking to her thanks to its lens. It is therefore no longer necessary to start your sentence with “Alexa!”. For speakers without a camera, proximity sensors come into play for the same purpose. “We merged data from an Echo’s sensors (the camera, voice input, its ability to detect presence) with AI models capable of understanding these non-verbal signals. Amazon specifies in its press release. “We also focused on reducing latency so that conversations flow naturally, without pauses, and responses are the right length for the voice, rather than listening to paragraph after paragraph read aloud“The second advantage of generative AI is that it takes context into account by relying on all its knowledge linked to the person who is calling on it.”The next generation of Alexa will be able to deliver unique experiences based on the preferences you’ve shared, the services you’ve interacted with, and information about your environment. Alexa also conveys relevant context throughout conversations, the same way humans do at any time“. In short, a “friend” always available and ready to discuss anything and everything and who, in addition, has personality, humor and… subjectivity. The Large Language Model (LLM ) for broad language model used for Alexa’s generative AI was tailor-made for Amazon.”Alexa will have a point of view, making conversations more engaging. Alexa can tell you which movies should have won an Oscar, celebrate with you when you answer a quiz question correctly, or write an enthusiastic note to send to congratulate a friend on their recent graduation” says Amazon.
This new generation Alexa should arrive in the United States before the end of the year and later in Europe. The good news: it will be able to integrate all Amazon Echo devices, even first generation ones.