AI Transparency Statement

We build AI systems, and we use AI in our own work. This statement says exactly where, what it decides, and what it does not. It exists because you are entitled to know when you are dealing with a machine and when a person is accountable.

1. The Short Version

• No decision about a person is made by AI alone. Not in recruitment, not in credit or billing, not anywhere else.
• We do not use automated systems to reject, rank, or score job applicants. Every screening decision is made by a person who can explain it.
• Where you interact with an AI system we build or operate, you will be told at the start of the interaction, not in a footnote.
• Client data is never used to train third-party models. We configure enterprise endpoints with training disabled.
• A named person holds editorial responsibility for anything we publish, whether or not AI helped produce it.

2. Where AI Appears in Our Services

AI Consulting and Systems Development

We design and build AI systems for clients. In those engagements the client is the deployer and is responsible for how the system is used; we are responsible for building it to the agreed specification, including the evaluation, guardrails, logging, and human oversight points set out in the statement of work. We will not build a system that makes legally significant decisions about individuals without a human review step, and we will say so during scoping rather than at handover.

Chatbot and Conversational Design

Assistants we build identify themselves as automated at the start of a conversation. They are scoped to a defined set of tasks, refuse questions outside that scope rather than guessing, and offer a route to a human at any point. High-risk intents — anything involving money, safety, health, legal rights, or account closure — are designed to route to a person by default.

Recruitment

This is the area where the stakes for individuals are highest, so the rules are strictest:

• AI is used to help find and organise candidates. It is not used to decide about them.
• No candidate is rejected, ranked, scored, or filtered out by an automated system.
• Every technical screening assessment is carried out by a practising engineer, who writes and signs the assessment you receive.
• We do not use video interview analysis, emotion or affect inference, personality inference, or any tool that attempts to infer characteristics from voice, face, or writing style. These techniques have a poor evidence base and, in the case of emotion inference in an employment context, are prohibited under the EU AI Act.
• Candidates may ask how a decision about them was reached, ask for a review by a different person, put their own case, and contest the outcome.

This is a deliberate commercial choice. Automated screening would be cheaper for us. It is also the single most legally exposed and least reliable thing we could do with this technology, and we are not willing to do it to the people whose careers are involved.

Fashion Design

We may use AI tools for early concept exploration and mood boarding. Technical drawings, tech packs, patterns, and specifications are produced and verified by a human designer, because a factory builds from those documents and an error in them is expensive in cloth. Where an AI tool contributed to a concept, we will tell you if you ask.

3. Where AI Appears in Our Own Work

• Software development: AI coding assistants help write and review code. All code is reviewed by an engineer, tested, and covered by the same warranties as any other code we write.
• Written material: AI may assist with drafting and editing. A named person reviews everything we publish and holds editorial responsibility for it. We do not publish AI-generated text that has not been reviewed by someone who can defend its accuracy.
• Images: we do not publish synthetic photographic imagery that could be mistaken for a real person, place, or event. Where we use generated illustration, it is visibly illustrative.
• Internal analysis: AI helps us summarise and organise information. It does not make hiring, pricing, or client decisions.

4. What Happens to Data We Put Through AI Systems

• Training is disabled. We use enterprise endpoints configured so that your content is not used to train the provider's models, and we verify that configuration rather than assuming it.
• Retention is minimised. Zero or minimal retention is configured wherever the provider offers it.
• Providers are named. Model providers used in your engagement are identified in the statement of work and listed on our sub-processors page.
• Special category data is not submitted to a model provider without your express written instruction and a documented assessment.
• Where an engagement requires it, we can run open-weight models entirely within your own infrastructure so that no content leaves your environment at all.

5. What AI Cannot Do Reliably, and What We Do About It

We would rather set expectations honestly than oversell the technology:

• Language models can produce confident, fluent, incorrect output. We manage this with retrieval grounded in your own sources, explicit refusal behaviour when confidence is low, scope enforced outside the prompt, and evaluation suites that test for exactly this before each release. We do not claim to eliminate it.
• Quality drifts. Model updates, changing source content, and shifting user behaviour all degrade a system that is not monitored. This is why we build evaluation harnesses before we build features.
• Bias is not solved by good intentions. Where a system's output could affect people differently by group, we test for it, document what we tested, and report what we found — including when we found something.

6. Your Rights

If an AI system we operate has been involved in a decision affecting you, you may: ask whether AI was involved and how; ask for a meaningful explanation of the logic and its significance; request review by a person; express your point of view; and contest the outcome. Write to privacy@etradersgroup.online. We respond within one month, and there is no charge.

If the system belongs to one of our clients rather than to us, we will tell you who the responsible organisation is and pass your request to them.

7. Regulatory Position

We monitor the EU AI Act, which classifies AI used for recruitment, candidate filtering, and employment decisions as high-risk, and which requires disclosure when a person interacts with an AI system. Because we have chosen not to use automated systems to make decisions about candidates, we do not place ourselves in the high-risk provider category for our recruitment service. Where we build AI systems for clients, we identify during scoping whether the intended use is high-risk, tell the client plainly if it is, and document the obligations that follow — including the fact that putting their own name on a system can make them its provider in law.

We also comply with Article 22 of the UK and EU GDPR on automated decision-making, as described in our Privacy Policy.

8. Keeping This Current

This is a fast-moving area in both technology and law. We review this statement at least every six months and whenever we materially change how we use AI. The version and last-reviewed date are at the top of this page. If you believe something here is inaccurate or out of date, please tell us — we would rather be corrected than be wrong.