Table of Contents
Spotify has broadened its collaboration with Google Cloud to leverage the capabilities of large language models (LLMs) in order to analyze and recognize the listening preferences of its users across podcasts and audiobooks.
On Thursday, Spotify announced an expansion of its collaboration with Google Cloud. This partnership aims to utilize large language models (LLMs) to analyze and understand users’ listening habits across podcasts and audiobooks. By doing so, Spotify can provide personalized recommendations that cater to individual preferences.
It has recently partnered with Google Cloud to utilize its LLMs (large language models) in analyzing its vast content library consisting of approximately 5 million podcasts and 350,000 audiobooks. The aim is to enhance the metadata of these podcasts and audiobooks, as stated in a press release.
Metadata includes crucial information such as the title, author or hostname, show notes, and other details that appear in search results on podcast platforms like Spotify and Apple Podcasts. The Verge has contacted Spotify to clarify the exact meaning of “augment” in this context.
Advancements in AI: Spotify’s LLM Implementation
LLMs like OpenAI’s ChatGPT and Google Bard utilize artificial intelligence to generate text and other content by being trained on an extensive volume of data.
Google Cloud, a subsidiary of Alphabet, possesses multiple LLMs including PaLM 2, Codey, Imagen, and Chirp. These LLMs undergo training on various data types such as text, codes, images, audio, and video.
Spotify was one of the pioneers in utilizing AI for music recommendation algorithms, a decade ago. Presently, the Swedish firm is looking to implement LLMs to extend this technology to its non-music content, including podcasts and audiobooks.
The music streaming company has been aiming to enhance its profits by expanding its range of revenue-generating mediums, including podcasts and audiobooks.
Enhancing Spotify’s Content Curation with Google AI
Identifying “harmful” content appears to be another aim of its Google Cloud AI experiment, though the company didn’t specify much about what that will entail. The platform’s rules for podcasters and musicians ban several types of “sensitive” and “dangerous” content — and it currently uses a mix of automated technology and human reviewers to enforce these rules.
The Verge has reached out to its music web for more details and will update you when we hear back.
Spotify’s personalized recommendations for podcasts and audiobooks will be enhanced with the help of Google Cloud’s AI tools. The LLMs will be utilized to gain a deeper understanding of the patterns behind users’ preferred spoken content, leading to more customized recommendations.
At present, it provides recommendations for new podcasts and specific episodes to its listeners through its home screen. Additionally, each podcast’s show page includes a “More like this section” that offers similar podcasts. However, these features may not always be accurate, as several users have pointed out.
Spotify’s AI Expansion and Content Diversification
For instance, its suggestions for Overheard at National Geographic include a podcast about nursing, one on bipolar disorder, and a true crime podcast.
It introduced an AI tool called AI DJ earlier this year, which utilizes OpenAI’s generative AI tools to produce personalized playlists. The feature has received mixed reviews thus far. Although Spotify’s use of AI for spoken audio is not a standalone feature like AI DJ, it demonstrates the company’s commitment to enhancing the user experience for its non-music services.
The company had previously made assurances of generating lucrative profits through its expensive venture into podcasts and audiobooks, emphasizing high-margin returns.
Gustav Söderström, the chief product and technology officer of Spotify, acknowledged that Google Cloud has been dedicated to creating an exceptional platform for their products to operate on, and has kept pace with the advancements in technology by exploring the potential of generative AI to drive innovation.
It is delving into the use of LLMs to ensure a secure listening experience and detect any harmful content, in addition to its extended collaboration with Google.
