Suno

Suno generates a complete song — lyrics, melody, sung vocals, and instrumental arrangement — from a simple text or a few keywords. The user needs no musical skills whatsoever.

Two main modes: “Describe Your Song” (a few words suffice, Suno generates lyrics and music) and “Custom” (the user writes their own lyrics and selects the genre). In both cases, Suno produces multiple variations that the user can listen to, compare, and refine.

The result is immediate (30 to 60 seconds) and requires no knowledge of music theory, composition, or production. This radical accessibility is what makes it a potential mediation tool.

Unlike ChatGPT, Suno is not spontaneously used by patients for emotional support. Its use in mental health is emerging and clinician-led, primarily driven by practitioners integrating it into creative mediation frameworks. The scientific literature on Suno specifically is still very limited, but two types of data exist.

In occupational therapy, mediation is the third element in the therapeutic relationship: clay, wicker, drawing serve as concrete supports that mediate the exchange between patient and clinician.

Gaëlle Charlot proposes the concept of “non-perceptible third” to describe what generative AI does within this framework: a third element that is neither a physical material nor a conventional piece of software, but an unpredictable semantic transformation. We don’t know exactly how Suno interprets the patient’s words, and it is precisely this opacity that creates a therapeutic space: the result surprises, displaces, and opens up dialogue.

This concept echoes Winnicott’s transitional object (neither fully internal nor fully external) and Pease’s therapeutic computational creativity FR (2022) — the idea that AI’s generativity can be therapeutically productive precisely because it escapes the control of both the patient and the clinician.

Suno represents a different case from ChatGPT. It is not a tool that patients spontaneously use to talk about their problems — it is a creative mediator integrated into a therapeutic framework by a professional. The distinction is fundamental.

What makes Suno clinically interesting is the combination of three properties: radical accessibility (no prior skills required), immediacy of result (a complete song in 30 seconds), and controlled variation (the same text in multiple styles). These three elements allow reaching patients who are typically failed by traditional mediations: cognitive impairments, lack of education, frustration intolerance.

The testimony of the poly-substance use disorder patient who discovers an identity as a “creator” is particularly striking. This is a shift from pleasure through substance to pleasure through creation — which, in the addiction field, constitutes a significant marker of recovery.

Caution is warranted, however: we are at the case report stage, not at evidence level. Gaëlle Charlot’s protocol is led by an experienced practitioner, within a specific institutional setting. Its transfer to other contexts will require formalization and rigorous evaluation. But the signal is interesting enough to merit attention from clinicians and researchers in digital art therapy.