Nvidia has unveiled Fugatto, a generative audio AI model that can create sounds, music, and voices from text and audio prompts.

What sets Fugatto apart is its ability to modify audio characteristics, such as changing accents and emotional tones in voice recordings.

Despite its advanced capabilities, CEO Jensen Huang stated that fully trustworthy AI systems are still several years away, and Fugatto is currently not available for public use.

Bryan Catanzaro, Nvidia's VP of applied deep learning research, highlighted that Fugatto is versatile, performing multiple tasks that typically require separate models.

However, Catanzaro also emphasized the risks associated with generative technology, indicating that Nvidia is carefully considering how to manage these concerns.

Fugatto joins a competitive landscape of generative AI tools, with similar offerings from companies like Meta and Google, which are also exploring audio generation technologies.

Nvidia has faced scrutiny over its AI training practices, particularly regarding the use of subtitles from thousands of YouTube videos.

This scrutiny aligns with broader concerns in the music industry, where artists have called for an end to the use of their works in AI training without proper compensation.

As generative AI continues to evolve, the music industry is increasingly anxious about copyright infringements and potential job losses for voice actors.

Nvidia's growth trajectory is complicated by challenges such as production issues, customer concentration, and geopolitical risks, including potential tariffs.

Despite these challenges, Nvidia's stock has surged over 210% this year, reflecting strong investor confidence in the company's AI technologies.

The development of Fugatto, which took over a year to enhance its creativity without significantly increasing the dataset size, showcases Nvidia's commitment to innovation in AI.