OpenAI aims to use AI to "create music," intensifying competition with Google and startups

After the launch of the text-to-video model Sora caused a sensation, OpenAI is now turning its attention to a new creative field: artificial intelligence music generation.

According to a report by technology media The Information on the 24th, OpenAI's internal team has been taking steps to develop music generation AI. Another insider indicated that the company is collaborating with some students from the Juilliard School to annotate musical scores, which is necessary work to provide training data for developing music generation AI.

These developments suggest that OpenAI is accelerating its product expansion efforts to extend the daily usage time of its over 800 million users and explore new revenue growth points. This move will put OpenAI in direct competition with Google, which has already launched similar products, as well as emerging startups in the AI music field like Suno and Udio, with copyright issues being a key challenge all participants must face.

Currently, the launch time of OpenAI's music generation plan, as well as whether it will be released as a standalone application or integrated into ChatGPT or Sora, remains unclear. An OpenAI spokesperson declined to comment on this.

Expanding Product Lines, Targeting Google and Emerging Unicorns

OpenAI's entry will directly intensify competition in the AI music generation field. One of its main competitors is tech giant Google. Google launched its second-generation music production model Lyria in May this year, offering services to customers through Google Cloud and promoting its capabilities in advertising music composition to marketers.

At the same time, OpenAI will also face challenges from rapidly growing specialized AI music startups like Suno and Udio. Founded three years ago, Suno has achieved an annual recurring revenue of approximately $150 million by selling subscription services for its AI music generator, nearly quadrupling compared to a year ago. Launching an AI music model capable of generating vocal tracks or instrumentals would allow OpenAI's functionality to surpass the current limitations of ChatGPT, which can only generate text content such as lyrics and chords.

OpenAI's latest initiative is a continuation of its strategy to go beyond its core chatbot service and build a more diversified product ecosystem. Following the tremendous success of ChatGPT, the company has been actively exploring new application scenarios to solidify its market leadership and create new commercial value.

This move closely follows the release of its video application Sora. Sora allows users to generate TikTok-style AI short videos, and according to its head Bill Peebles, the application achieved one million downloads within less than five days of launch, growing even faster than ChatGPT, which debuted at the end of 2022. Additionally, OpenAI is developing a social media-like information flow feature within its chatbot to provide potential channels for users to share their created AI music. The integration of these social features aims to further enhance user engagement

Potential Applications and Business Models

OpenAI's exploration of music-generating AI showcases its vast imagination for commercial applications. According to reports citing a third party involved in related discussions, the company has internally discussed the possibility of generating music through text and audio prompts, such as having AI add guitar accompaniment to existing vocal tracks.

Such products can not only help ordinary users easily score music for video content but also provide efficient creative tools for enterprise clients in industries like advertising. For example, advertising companies could use OpenAI's tools for lyric brainstorming, creating engaging ad songs based on music samples, or even uploading videos to mimic their style for creation. This could pave the way for OpenAI to expand its advertising business in the future.

However, to enter the AI music field, OpenAI may need to reach agreements with major record labels to avoid potential copyright lawsuits. This is a significant obstacle that all AI music generators cannot avoid.

The Recording Industry Association of America (RIAA), representing Universal Music Group, Sony, and Warner Bros., has already filed lawsuits against Suno and Udio, accusing the two startups of using copyrighted songs to train their AI models without permission or proper compensation, and seeking up to $150,000 in damages for each infringing song. In response, both Suno and Udio stated that they could use copyrighted materials under the "fair use" principle in copyright law.

As reported earlier this month, Universal Music and Warner Music are negotiating AI licensing agreements with these startups and Google. To mitigate risks, OpenAI has taken precautions in its existing products, such as ChatGPT not sharing the full lyrics of certain songs. Its CEO Sam Altman also stated after the release of Sora that they would give copyright holders "more refined control over character generation" and share a portion of video generation revenue with them