AI Disclaimer

1. What AI features do

Depending on the feature and configuration, RoboSelf may use third-party or user-selected models to summarize market information, explain metrics and indicators, answer questions, draft or modify strategy rules, generate code or structured configurations, review strategy behaviour, rank or optimize candidates, debug logic, and help design or operate an AI agent.

Some Service outputs are deterministic calculations produced by conventional code; others are probabilistic text or structured output from a generative model. A polished interface or structured response does not make model-generated content a verified calculation. Check the feature label, source data, methodology, and resulting configuration.

2. Do not rely on AI as advice or a suitability decision

AI output is presented for informational and software-assistance purposes. The model does not know your complete financial situation, objectives, obligations, or ability to bear loss. Do not rely on AI output as personalized financial, investment, legal, tax, accounting, fiduciary, suitability, or brokerage advice, or assume that directive language is a suitable recommendation to buy, sell, hold, rebalance, or trade any instrument or strategy.

A model may produce directive language, predictions, ratings, or apparent recommendations. Treat that as an error-prone generated response—not as a suitability determination or sufficient reason to trade. Consult qualified professionals for advice specific to you. Nothing in this disclaimer excludes a status or duty that applicable law imposes based on the Service's actual conduct.

3. Hallucinations, inaccuracies, and missing context

AI may:

• invent prices, events, citations, filings, news, calculations, features, broker capabilities, or legal rules;
• use stale, delayed, incomplete, biased, or incorrectly mapped data;
• misread a chart, table, prompt, strategy, time frame, unit, symbol, market session, or account state;
• omit a material risk, exception, fee, constraint, counterexample, or conflicting fact;
• produce different answers to the same question or contradict itself within one response; and
• state uncertainty poorly or express false confidence.

Verify factual claims against current primary sources, broker records, and the actual strategy configuration. Do not rely on a citation until you have opened and checked the cited source.

4. Strategy, code, and optimization risk

AI-generated rules, parameters, code, and configurations can contain syntax errors, unsupported fields, insecure logic, look-ahead bias, unintended order semantics, missing exits, excessive risk, contradictory conditions, or assumptions that cannot be executed. A generated strategy can lose some or all capital even when its explanation sounds reasonable.

AI-generated or AI-selected optimization results can overfit historical noise. A backtest that runs successfully proves only that the configuration ran against the selected data and assumptions; it does not prove correctness, robustness, suitability, or future profitability. Review the Trading Risk Disclosure.

5. AI agents and actions

An AI agent may observe, summarize, recommend, request approval, or, in specifically enabled contexts, take an eligible automated action. The permission label and saved policy—not an agent’s prose—determine what the system may attempt. Model output can still select an incorrect action or rely on incorrect inputs.

Review every selected input, action permission, confidence threshold, model-fallback setting, schedule, strategy assignment, and safeguard. Start with observe-only or paper modes, inspect logs, and use the least authority necessary. “Approval required,” “shadow,” “confidence,” and other safeguards reduce but do not eliminate risk. If autonomous activity is enabled, individual actions may occur without a new prompt.

6. Human review is required

Before using AI output, you are responsible for:

• checking the source data, timestamp, symbols, assumptions, math, and factual claims;
• comparing the generated output with the configuration actually saved by the Service;
• reviewing code and rules for security, validity, supported behaviour, and unintended side effects;
• testing with independent data and paper trading while accounting for simulation limits;
• deciding whether the output is lawful and appropriate for your own use; and
• continuously monitoring any strategy or agent that you deploy.

Do not use AI as the sole control for a time-sensitive, high-impact, or irreversible decision. If you cannot independently evaluate an output, do not apply it to a live account.

7. Model and provider limitations

Model availability, knowledge, behaviour, context limits, safety filters, pricing, latency, and output format can change. A selected model may be unavailable and, if you enabled fallback, another model may be used. Providers can reject or truncate requests, return malformed data, experience outages, or change their terms and retention practices.

RoboSelf does not train or control the underlying third-party models and cannot guarantee how they learned a fact or why they produced an output. Model names and provider descriptions identify routing or configuration; they are not warranties of capability, recency, or suitability.

8. Data, privacy, and confidential information

A prompt may be sent with the market, strategy, portfolio, or account context selected or required by the feature. It may pass through an AI routing provider to the chosen model provider. If you supply your own API key or provider account, the provider’s terms and settings may independently govern the request.

Do not enter broker passwords, private keys, full payment-card information, authentication secrets, material non-public information, another person’s personal information, or confidential material you are not authorized to disclose. Review the Privacy Policy and the selected provider’s terms before use.

9. Bias and inappropriate content

Models can reflect bias in training data, provider policies, prompts, and source selection. They can produce offensive, discriminatory, misleading, or otherwise inappropriate content. Report problematic output and do not publish or act on it without review. You remain responsible for content you choose to use or share.

10. Intellectual property and third-party rights

AI output may resemble existing content, include protected material, or fail to qualify for copyright or other protection. RoboSelf does not promise that output is unique, accurate as to attribution, non-infringing, or exclusively owned by you. You must review outputs, comply with provider terms, and obtain any permission needed before publishing, licensing, or commercializing them.

11. No guarantee

RoboSelf does not guarantee AI accuracy, completeness, recency, availability, explainability, consistency, security, originality, suitability, regulatory compliance, or profitability. AI output cannot predict future market performance or guarantee that a risk control or strategy will work.

The warranty disclaimers and liability limits in the Terms of Service apply. If you do not accept these limitations, do not use AI-enabled features.