AI Policy

​

This addendum forms part of the Counselling Agreement and Privacy Information. It explains how I may use artificial intelligence (AI) and other digital tools in my practice, and sets out considerations for your own use of AI alongside therapy. We will discuss it before any AI tool is used with your personal information.

​

AI is readily available on multiple platforms as part of everyday modern living. Whilst its appearance can be subtle, awareness and protecting your data is at the forefront of my focus when using digital systems. Therefore, this document may seem long and detailed, but is generated for full transparency to enable you to give informed consent.

​

What I mean by AI

AI includes generative AI chatbots and writing tools, automated transcription or note-taking, summarising and translation tools, and systems that analyse information or make suggestions. Digital services may also collect or monitor information even when they are not described as AI.

 

My use of AI and digital tools

I remain responsible for the service, my professional decisions, my conduct and the records I keep. I will not delegate clinical judgement to AI. I will not use AI to make a diagnosis, assess your risk, decide whether to continue therapy, or determine your treatment without my own independent professional assessment. Any AI-generated content that I choose to use will be critically pre-reviewed by me for accuracy, relevance, bias and potential harm.
 

Before using any AI or digital tool in a way that involves your personal information, I will explain the tool, its purpose, what information would be entered or collected, how the information is handled and stored, relevant risks, and any available alternative. I will obtain your informed consent before entering your personal information into an AI tool. I will not enter identifiable or confidential information about you into a public or consumer AI service such as a chatbot.
 

Consent to this use is optional and separate from your agreement to receive counselling. You can say no or withdraw consent for future use without affecting your access to therapy. I will continue to meet my separate data-protection obligations; consent does not replace the need for a lawful basis or other safeguards. I have completed an audit of AI use in my business and review it at least annually. This audit supports, but does not replace, checking each proposed tool and use for competence, data handling, confidentiality risks and safeguards before it is used with client information. I will review the audit sooner if a tool, purpose or data-handling arrangement changes.
 

I am of total reluctance and will not record, transcribe or upload a session, message, document or other client material to an AI service unless the particular purpose, tool, information, risks, storage and retention have been explained to you and you have given specific informed consent in advance. I do not use AI with clients’ personal information. I will not use your personal information to train or improve an AI model. If my proposed use changes, I will explain the change and seek fresh consent before proceeding.

I remain responsible for all professional decisions and review any AI output myself.

To manage my practice effectively, maintain clinical standards, and fulfil my professional obligations, I process information categorised as Special Category Data under the UK GDPR. Specifically, this encompasses your mental and physical health details, clinical histories, intake assessments, and session summaries.
 

I process this sensitive personal data relying concurrently on two strict legal bases under the UK GDPR:
• Your Explicit Consent pursuant to Article 9(2)(a).
• The Provision of Health or Social Care Treatment pursuant to Article 9(2)(h),
subject to strict professional duties of confidentiality.

If I use any AI or digital tool with your personal information, I will complete the details below  and make sure they match my privacy notice. If I do not use such tools, I will state that clearly.

​

Provider(s) & Tools:
 

Practice Management and Digital Systems

WriteUpp (Clinical Records & AI Scribe)

I utilise the specialist healthcare practice management platform WriteUpp to schedule appointments and host clinical notes. Data is fully encrypted within secure single-tenant structures, is never exposed to public repositories, and is never used to train third-party commercial large language models (LLMs).

 

Tide Banking (Financial Administration & Business Banking Accounts)

I utilise Tide Platform Limited to manage business banking and receive payments. When you settle invoices, your payment metadata (including name, sort code, account number, and transaction references) is processed by Tide. Tide employs closed-loop, predictive machine learning algorithms strictly for real-time fraud monitoring, compliance screening, and network security. Additionally, Tide uses generative internal tools (such as Member Summarise and AgentAssist via AWS/Amazon Bedrock) which allow customer support staff to securely review transaction histories if I contact them. Your financial or medical data is completely isolated, never exposed to public repositories, and never used to train third-party AI models.

​

Xero (Invoicing & Financial Accounting)

For invoicing and administrative accounting, I utilise Xero. Your invoices necessarily include your name, the service and transaction totals, which connect directly to your physical and mental health care. Xero uses closed-loop, internal machine learning tools exclusively to parse uploaded invoices, auto-populate bookkeeping entries, and match incoming bank transfers. Your sensitive identifiers are completely isolated within my private, encrypted business account and are never used to train public foundational AI systems.

Wix (Website Hosting, Enquiries & Registration Forms)

My practice website is hosted on the Wix platform, which serves as the primary gateway for my client registrations and enquiry forms. When you submit an online booking, intake form, or enquiry, Wix secure databases process and store your initial mental and physical health summaries. While Wix utilises advanced back-end data architectures and infrastructure security protocols, your submitted form text is strictly ring-fenced. Wix's integrated AI configurations do not access, process, or retain your clinical intake fields or contact information to train any commercial generative models, ensuring your data remains secure and confidential from initial point of entry.

​

Communication and Session hosting

Zoom

To conduct remote video consultations, I utilise Zoom. The Zoom AI Companion is disabled and would ordinarily generate automated session summaries and text transcriptions. In accordance with Zoom’s strict security standards, your live audio, video feeds, and text logs are encrypted in transit and at rest. Zoom maintains zero-retention parameters, guaranteeing that your special category data is never utilised to train external public commercial AI models. International transfers are safeguarded under standard contractual clauses.

​

Microsoft Teams

 

Alternatively, I may facilitate virtual sessions via Microsoft Teams. I have disabled Microsoft Copilot / Teams AI features to draft speech-to-text summaries and logs. Ordinarily this would be protected by Microsoft’s Enterprise Data Protection (EDP) framework, all stream content and prompts are strictly bound inside my private business tenant. Your mental and physical health details are completely ring-fenced and never leave this boundary to train commercial foundational models.
 

Supporting business operations & productivity

​

GoodNotes

For immediate handwritten notations, I utilise GoodNotes on a secure tablet. GoodNotes employs localized, on-device machine learning algorithms solely to translate handwriting into searchable digital text. Cloud-based generative features are disabled, preventing your sensitive notes from interacting with external networks or servers.

​​

Gmail & Microsoft 365 (Administrative Suite)

General business operations and correspondence are conducted via Gmail and Microsoft 365. Machine learning within these environments is confined strictly to foundational, automated administrative security infrastructure, including anti-phishing triggers, spam screening, and automated calendar scheduling. No email text or health information is harvested for commercial language model training.

 

Canva

I utilise Canva exclusively as a visual layout tool to design generalized marketing assets and public informational resources. No client personal identifiers, clinical files, or physical/mental health details ever touch or interact with the Canva platform or its integrated AI design engines.

 

I access the previously listed services from the UK and have selected UK-based settings where available. This does not necessarily mean that all information is stored or processed only in the UK. Providers and their subprocessors may store, access or process information in other countries under their applicable terms and safeguards. I review each service’s relevant settings, privacy information and data-processing terms as part of my business AI audit, which is reviewed annually and when circumstances change.

​

Your use of AI alongside therapy

You may choose to use AI independently, including to reflect, organise thoughts or find general information. This is your choice and will not affect our work together. AI responses can be inaccurate, incomplete, biased or unsuitable for your circumstances, and an AI service may store or use what you enter under its own terms. Please avoid sharing your name, contact details, session notes, identifying details about other people, or other confidential information with an AI service unless you have checked and are comfortable with how that service handles it. AI does not operate with real empathy and only answers based on patterns of information entered into it and is likely programmed to provide user bias.

 

An AI tool is not a substitute for a therapist, GP, crisis service or emergency help. I cannot see or monitor your AI conversations and cannot verify what a tool tells you. If an AI interaction concerns you, or affects how you are feeling or what you want from therapy, you are welcome to bring that experience to a session. You do not have to disclose or show me AI conversations.

​

If we consider using AI together as part of therapy, we will first agree the purpose, the specific tool, what information will be entered, the risks and alternatives. We will not upload a session recording or identifying or confidential information about you or another person without your specific informed consent and the required privacy safeguards. You can decline or stop this activity at any time.