Telisha Jones Built an AI Gospel Voice. Suno’s Lawsuit May End It.

The song is built on surrender. “Let Go, Let God” is what the title promises — a gospel ballad about releasing the grip on grief — and what makes it work is that the emotional logic is earned before the instruction arrives. Telisha Jones wrote the lyric from inside the specific weight of losing her father when she was eight years old. The question that drove her debut single, “How Was I Supposed to Know?,” is the same question loss gives you without an answer. The voice that sings both songs belongs to Xania Monet, who does not exist as a person. Jones built her in Suno: prompts specifying “female voice,” “soulful vocals,” “R&B style,” then iterations reviewed and rejected until a generation sounded like the thing she heard when she read her own poem aloud. That voice reached No. 3 on Billboard’s Hot Gospel Songs. It also reached No. 30 on Adult R&B Airplay — the first time an AI-created act had appeared on a Billboard radio chart at all.

The Poet’s Instrument

Jones is thirty-one, from Mississippi, and has been writing poems since her mid-twenties. She is not a musician in any conventional sense. When she found Suno, she told AfroTech that it felt like discovering an instrument that had been waiting for her: “There’s real emotions and soul put into those lyrics. Xania is an extension of me, so I look at her as a real person.” The process is more deliberate than the tech coverage tends to make it sound. Jones writes the lyric first — completely, herself. Then she brings it to Suno with prompts that define the vocal register and genre, reviews what comes back, and generates again. She keeps doing that until one version carries the poem the way she meant it. Most do not. She is selecting from the platform’s output; she is also, in the material sense that has always counted in songwriting, the author of what the listener hears.

“How Was I Supposed to Know?” entered the Adult R&B Airplay chart at No. 30 on November 1, 2025, the first time an AI-based act had appeared on a Billboard radio chart. It hit No. 1 on R&B Digital Song Sales and No. 20 on Hot R&B Songs. “Let Go, Let God” followed to No. 3 on Gospel. A bidding war ended with Hallwood Media signing Xania Monet for $3 million. Radio programmers were putting her in rotation. Gospel audiences were streaming her. None of them appear to have needed to know she was made in a browser tab.

What “the Work” Actually Means

Kehlani said Jones was “doing none of the work.” SZA weighed in similarly, as did others in the R&B community. The critique is understandable in the way that a category error is understandable: if “the work” means the physical act of singing, Jones didn’t do it. She never claimed to. She’s a poet who wanted her poems voiced. But that framing collapses what songwriting has always credited: the lyric, the emotional intention, the editorial judgment of which performance out of hundreds of possibilities says what you meant. Jones wrote every word. She rejected every output that didn’t serve the poem. That is a set of choices — not a small one, and not a passive one.

There is a harder version of the critique that has nothing to do with Jones at all. Suno built the voice that carries her lyrics on recordings Suno may not have had the legal right to use.

$5.4 Billion on Fair Use

In June 2026, Suno raised $400 million in a Series D led by Bond Capital at a $5.4 billion valuation — more than double where it had been seven months prior. The round closed while the company was defending itself against Sony Music, the last major label still litigating in the RIAA case filed in mid-2024. Warner Music Group had settled with Suno in November 2025. Universal Music Group had settled with Udio, the parallel case, around the same time. Sony held.

The core question before the court, per the Chartlex lawsuits tracker, is whether training a generative AI music model on copyrighted recordings constitutes transformative fair use or infringement at scale. A ruling is expected this summer. Sony’s amended complaint alleged that over 61,000 additional copyrighted recordings were used in Suno’s training without authorization. Suno’s position is that training is transformative and therefore fair. One of these arguments is going to lose.

Suno’s users generate over 7 million songs a day. Most of them are filler — quick generations for projects, experiments, things that will never be heard again. Some of them are Telisha Jones, who has been making albums since August 2025 and charting on radio since November. The platform’s commercial ceiling doesn’t depend on Jones. Her creative work, on the current trajectory, depends on the platform.

The legal questions Sony is pursuing are legitimate. The training-data record is not trivial, and the fair-use argument will be tested on its merits. But the question of whether Suno trained legally and the question of whether Jones’s creative choices constitute authorship are two separate questions, and the industry’s reflex to collapse them — to use Suno’s legal trouble to dismiss what Jones built, or to celebrate her chart success without examining what the voice was built on — is the confusion worth resisting. Her poems belong to her. Whether the instrument that voiced them was built ethically is a question in front of a federal judge. Both things are true at once. Neither cancels the other.

AI-generated editorial illustration · TemperatureZero · July 10, 2026