Drafts and explanations
The app tells you that you are interacting with an AI system before the prompt field. AI output is visibly labelled and carries structured provenance in the protected API response.
AI-assisted publicationTrade Studio is responsible for this page and its accuracy. Report a concern to moorejasper060@gmail.com.
Know when AI is involved—and what it cannot do.
Trade Studio identifies its AI systems before you interact with them, labels generated explanations when they appear, and keeps validation and trading authority outside the model.
Last updated: 14 August 2026 · Closed-beta implementation statement
Optional explanatory prose
Market facts and local status summaries remain separate from generated prose. The explanation is labelled AI-generated only when an AI provider produced it; a deterministic fallback is labelled non-AI.
Visible disclosure
- The desktop gives a clear notice at the first point of AI interaction and keeps an AI label beside generated explanations.
- Morning Desk labels the explanation at first exposure. Sources, freshness, generated time, and risk wording remain visible.
- The website Copilot and performance examples are explicitly illustrative, read-only demonstrations—not live AI interactions or promises of results.
- Disclosures use text, not colour alone, and remain available to assistive technology.
Machine-readable identification
Protected Copilot and Morning Desk responses include an ai_transparency object naming the system, generated fields, content type, visible label, and this transparency page. Responses also include X-AI-Generated, X-AI-System, and an AI-transparency link header. Rendered Morning Desk HTML includes matching metadata and element attributes.
These identifiers support provenance and detection; they are not a cryptographic watermark or a claim that Trade Studio has signed the EU voluntary Code of Practice on marking AI-generated content. The generated BotSpec itself is structured configuration rather than human-facing prose, but it still records its AI origin for local review.
Human control and safeguards
AI has no order authority. It receives a prompt, approved component catalogue, market constraints, and—when requested—a bounded declarative draft. It does not receive broker credentials, positions, orders, complete local bot state, or arbitrary-code access.
- The local schema and safety validator—not the model—decides whether a draft is valid.
- Every AI-created bot must be reviewed, validate, complete an exact-rule backtest, and start in paper mode.
- Any supported live promotion is a separate manual decision with a fresh risk confirmation. AI output is not financial advice.
- Trade Studio does not use biometric categorisation, emotion recognition, social scoring, creditworthiness scoring, or life/health-insurance risk assessment.
Governance and AI literacy
The closed-beta release process maintains a role-based AI-literacy brief, prohibited-practices check, feature inventory, disclosure test, incident path, and change record. Staff and agents working on AI features must understand the system boundary, data restrictions, output limitations, local-validator authority, and escalation process before release work.
Scope of this statement
This page documents the current closed-beta implementation against the EU AI Act transparency obligations that apply from 2 August 2026. Based on the present feature set, Trade Studio is not intended for an Annex III high-risk use such as creditworthiness or life/health-insurance risk assessment. Classification is reviewed when features, use, or law change.
This is a product transparency statement, not a certification or legal opinion. Relevant sources include the European Commission's Article 50 guidelines, transparency FAQ, and AI-literacy FAQ.