AI‑generated complaints and how to deal with them


By Keara Bowgen-Nicholas | April 23, 2026

Last Updated on September 3, 2026

Artificial intelligence (AI) is becoming impossible to ignore in social care – not just in the tools providers use but in the correspondence they receive. Keara Bowgen-Nicholas, Senior Associate in the Dispute Resolution and Health & Social Care Regulatory teams at RWK Goodman, explores what that means for providers.

We are increasingly seeing AI-generated complaints that have clearly been drafted, polished or even entirely written with the help of AI tools. For complainants, especially those who feel overwhelmed, emotional or unsure how to articulate their concerns, AI lowers the barrier to making the complaint. It empowers residents, their families or friends to express themselves in a way that feels more confident and ‘official’. The Parliamentary and Health Service Ombudsman (PHSO) has reported a significant increase in generative-AI complaint submissions – a trend that is unlikely to slow down.

Tools such as ChatGPT, Gemini and others can help:

  • Draft, structure, refine and clarify a complaint.
  • Reframe emotional experiences into formal language.
  • Add references to legislation or Care Quality Commission (CQC) standards.
  • Make a complaint sound more authoritative.

How AI-generated complaints differ from traditional ones

While the growing use of AI is enabling people to feel more confident about raising complaints, it is presenting new challenges for providers. While providers might assume that AI-generated complaints would be clearer, we are seeing that, in practice, they often introduce additional and unnecessary complexity. The PHSO notes that AI-generated complaint text can become overly complicated, include irrelevant legal references, and may require resubmission – potentially increasing review time for everyone involved.

For example, AI ‘assisted’ complaints can:

  • Be longer and more repetitive: AI loves a list and really likes repeating itself. Providers may find themselves wading through lengthy documents that ultimately raise just one or a very limited number of issues.
  • Sound legalistic but are not always accurate: AI can churn out references to legislation, regulations and case law but these are not always reliable. AI has been known to ‘hallucinate’ (basically make things up), refer to out-of-date information, take sources out of context or even draw on how these work in other jurisdictions.
  • Appear more serious than the underlying issue actually is: A simple concern can suddenly arrive as a two-page letter filled with formal language. This can feel intimidating and may trigger overly defensive responses from staff.

Official guidance highlights these risks. The PHSO provides guidance on using AI prompts to help with complaints, warning that AI-assisted correspondence often adds complex legal phrasing and irrelevant procedural points that obscure the core issues. When key points are lost in generated text, resolving the complaint takes far longer. In short, AI is making complaints easier to generate but not always easier to interpret.

What does the law say about AI-generated complaints?

Legally, nothing changes – a complaint is a complaint, whether it has been made personally, by a family member, by a representative or produced with AI assistance. Providers’ duties under the Local Authority Social Services and National Health Service Complaints (England) Regulations 2009 remain the same: acknowledge the complaint, investigate it and respond within the stated timescale. The Information Commissioner’s Office takes a directly comparable position on AI-assisted requests, stating that “a request isn’t invalid just because someone’s used technology” – the substance is what matters.

Dismissing or minimising a complaint because it ‘sounds AI-generated’ would be procedurally unfair and is likely to escalate matters.

Common pitfalls providers should avoid

  • Under-engaging: Some providers instinctively dismiss AI-generated complaints as less ‘real’ or less valid, focusing more on the style than the substance. This risks missing genuine concerns.
  • Over-engaging: Others feel compelled to rebut every overly legalistic point. This leads to long, defensive responses that lose sight of what actually matters.
  • Adopting a defensive or cold tone: The formality of AI-generated complaints can cause providers to ‘mirror’ that tone. But formality is not a substitute for empathy, which could help de-escalate matters.

Behind every complaint, even an AI-polished one, it is important to remember that there is a person who is feeling distressed, frustrated or confused.

Practical tips

  • Have a strong, clear complaints policy and follow it: A policy should apply consistently, regardless of how the complaint was drafted. It should clearly outline where complaints should be sent, the investigation process and timescales for a response.
  • Acknowledge quickly: A prompt acknowledgement is one of the most effective ways to build trust, reduce anxiety and prevent escalation. Let the complainant know their complaint has been received and that it is being dealt with.
  • Keep communication flowing: If a deadline cannot be met, say so and explain why. Again, silence can escalate complaints.
  • Investigate carefully and forensically: AI-generated complaints often jumble or merge issues – do not be afraid to ask the complainant for clarification on the specific points they want addressed. Focus on facts, evidence and the remedy sought rather than writing style.
  • Respond clearly and proportionately: Keep responses calm and respectful. As the PHSO puts it, “simple, clear language that explains what happened will always be more helpful than complicated, formal writing” – that principle applies equally to providers’ responses.
  • Do not match the AI tone: If the complaint sounds like it has been drafted by a computer or robot, that does not mean the response has to. Use a natural professional tone – warm, clear and grounded.

Final thoughts

In practice, the key for providers is to be alert to the increasing use of AI in complaints while avoiding the temptation to treat them differently. Treat any AI detection score only as a triage signal, not proof of fraud. The CQC’s principles on AI in health and social care are a useful reminder here: effective governance and person-centred care remain the benchmark, whatever tools a complainant has used. By focusing on the substance rather than the style, and responding in line with established policies and procedures, providers can ensure complaints are handled consistently, fairly and proportionately.


Keara Bowgen-Nicholas is a Senior Associate in the Dispute Resolution and Health & Social Care Regulatory teams at RWK Goodman, where she advises health and social care providers on CQC compliance, investigations and complaints.

About Keara Bowgen-Nicholas

Keara Bowgen-Nicholas is a senior associate at RWK Goodman, specialising in data protection and information governance as well as inquests and regulatory matters for adult social care providers. She advises organisations facing investigations and enforcement action by regulators including the CQC/CIW and Ofsted and supports providers through inspections, compliance action and funding disputes.