The document from the Royal College of Speech and Language Therapists (RCSLT) sets out a framework for the safe, ethical, and effective use of artificial intelligence (AI) in speech and language therapy (SLT). It is aimed at helping therapists, employers, educators, and developers use AI responsibly while protecting patients and professional standards.
Here is a summary of the main points:
Purpose of the guidance
RCSLT recognises that AI is becoming increasingly relevant in healthcare and SLT practice. The guidance aims to:
- encourage innovation,
- improve access and efficiency,
- reduce administrative burden,
- and ensure AI is used in ways that are clinically safe, equitable, and person-centred.
Core principles
The guidance is built around several key principles:
1. Human oversight must remain central
AI should support — not replace — professional judgement. Speech and language therapists remain accountable for clinical decisions and outcomes.
2. Person-centred care comes first
Any AI system should enhance care quality and preserve therapeutic relationships rather than undermine them.
3. Safety and evidence are essential
AI tools should only be used when there is appropriate evidence of safety, effectiveness, and suitability for the intended population.
4. Transparency matters
Clinicians and service users should understand:
- when AI is being used,
- what it is doing,
- and its limitations.
5. Equity and inclusion
AI systems can introduce bias or worsen inequalities. Therapists should consider:
- accessibility,
- representation in training data,
- cultural and linguistic appropriateness,
- and impacts on underserved groups.
6. Data protection and confidentiality
Any AI use must comply with:
- GDPR,
- NHS/local governance rules,
- and professional confidentiality obligations.
Uploading identifiable client information into public AI systems without approval would generally be inappropriate.
7. Professional competence
SLTs are encouraged to develop “AI literacy” so they can critically evaluate tools, understand risks, and use systems responsibly.
Examples of potential AI use in SLT
The document discusses areas where AI may assist, including:
- clinical documentation,
- report drafting,
- transcription,
- administrative support,
- analysing speech/language data,
- and potentially augmenting assessment or therapy tools.
However, the guidance stresses that these uses require careful evaluation and governance.
Risks highlighted
The guidance warns about:
- hallucinated or inaccurate outputs,
- embedded bias,
- overreliance on automation,
- confidentiality breaches,
- reduced clinical reasoning,
- and erosion of trust if AI is poorly implemented.
Expectations for organisations and developers
RCSLT also calls on employers, commissioners, and technology developers to:
- involve SLTs and service users in design,
- provide training and governance,
- ensure robust testing,
- and maintain clear accountability structures.
Overall message
The overall tone is cautiously supportive:
- AI has genuine potential to improve SLT services,
- but adoption must be deliberate, evidence-based, and ethically governed,
- with clinicians remaining responsible for care quality and decision-making.
The Full guidance can be found here : Royal College of Speech and Language Therapists (RCSLT)