Amazon has unveiled a new AI-powered search feature for its Fire TV devices, aiming to improve the search experience for users. This new functionality leverages a custom-designed large language model (LLM) to enable more conversational interactions with Alexa, Amazon’s voice-controlled assistant.
Advanced Search Functionality
Fire TV users can now interact with their devices using natural language to find content more efficiently. Queries like “show British crime dramas with female leads” or “find the show about money laundering set in the mountains” are understood and processed by the large language model, facilitating a more intuitive search through the extensive libraries of streaming options.
This AI-enhanced search is being rolled out to Fire TV devices that are running Fire OS 6 or later. It can identify content based on various attributes such as topic, genre, plot details, actors, and quotes, utilizing data from services like IMDb. At a recent demonstration at Amazon’s Day 1 headquarters in Seattle, Joshua Park, senior product manager of Fire TV, demonstrated how the AI can manage specific requests.
Current Limitations and Future Development
Despite its promise, the AI search feature is not flawless. It encounters difficulties with broader queries and struggles after two follow-up questions. For instance, it had trouble accurately responding to the request, “show me Oscar-winning movies from the 1970s.”
Daniel Rausch, VP of Alexa and Fire TV, acknowledged these constraints and mentioned plans to enhance conversational capabilities in future updates. The goal is to enable users to refine their searches through an ongoing dialogue with Alexa, similar to getting recommendations from a knowledgeable store clerk. Currently, the feature’s ability to offer personalized, high-quality recommendations is still developing.
User Experience and Feedback
In practical use, the AI search has yielded varied results. While it understands conversational inputs, the quality of its recommendations can be inconsistent. For example, when asked to “show me some dark comedies with violence,” it suggested a mix of older films and unrelated titles. Similarly, a request for “TV series with more than six episodes that are highly rated” returned only two anime series, one of which had an average rating.
Amazon is working on improving these aspects. Park emphasized that the company is focused on enhancing the AI’s capability to accurately find content that meets users’ specific requests.
Additional Features and Future Prospects
Unveiled at Amazon’s annual devices event last September, the AI-powered search feature is part of a broader initiative to integrate generative AI into the Fire TV platform. Previously, Amazon introduced an AI-powered image generator for TV backgrounds, along with three Alexa experiences for playing games and creating songs.
Additionally, Alexa will be able to recommend titles included in users’ existing streaming subscriptions. For example, if a user asks for psychological thrillers with surprise endings, the results will be limited to content available through their current subscriptions. The new search capability is anticipated to be accessible in the coming weeks for U.S. customers with Fire TV devices running Fire OS 6 or higher.
Last Updated on November 7, 2024 7:58 pm CET
