Governance & AI Policy

• Every AI-generated or AI-synthesised block carries a visible AI-generated label together with the method used to produce it.
• Factual claims are traceable to the weekly source report they were extracted from; market figures shown on valuation cards are derived from public market data feeds and are clearly attributed on the card.
• Every AI block offers a Report this output link, which opens a message to info@ikaruscareer.com. Reports are triaged within five working days; content found to be materially wrong is corrected or withdrawn.
• Method: a weekly economist report is parsed by a large language model into structured fields (sentiment, confidence, mentions, valuation status). Scores are computed arithmetically from those fields and from historical price data — the model does not invent prices.
• Uncertainty & limitations: extraction can misread or omit content; sentiment labels compress nuance; confidence scores reflect the source commentary, not a probability of any outcome; coverage is limited to the assets discussed in that week's report.
• No AI output is used to make automated decisions about any individual.

• Prompts sent to our AI providers are constrained to macroeconomic extraction tasks; free-form user prompting of the model is not exposed in the product.
• Output controls: generated content is rejected or held if it contains personal recommendations to buy or sell, guaranteed-return claims, or harmful, deceptive, discriminatory, sexual, self-harm, or malicious material.
• Escalation: reports reach the operations owner immediately; suspected safety incidents are acknowledged within 24 hours, contained (content withdrawn or feature disabled), investigated, and followed by a written post-incident note. Affected users and, where legally required, regulators are notified.
• If you or someone else is at immediate risk of harm, contact your local emergency services — this platform is not a crisis service.

Today the only user contribution is an anonymous bullish / neutral / bearish vote. Individual votes are never published with your name, and there is no public posting, messaging, upload, or mentor feature.

If and when public posting, direct messaging, uploads, or mentor interaction are introduced, they will launch with: automated and human content filtering, in-product reporting on every item, user blocking and muting, community guidelines inside the Terms, published operator contact details, and active moderation with defined review times and an appeals route. Vote manipulation, brigading, and automated voting are prohibited and lead to suspension.

Report abuse or manipulation to info@ikaruscareer.com.

• Inputs: valuation scores use current price, 1/3/5-year average prices, and volatility from public market data, combined with the valuation stance extracted from the weekly report. Archetype accuracy uses that archetype's past recorded calls against later observed direction.
• Method: scores are deterministic formulas, published in outline on this page, not opaque model judgements.
• Uncertainty: historical averages are not forecasts; short data histories, illiquid proxies and ETF substitutes reduce accuracy; scores can move sharply week to week.
• Limitations: no personal circumstances, risk tolerance, tax position, or objectives are considered. Scores are never a recommendation.
• Correcting your data: your profile holds only your name, email, and avatar from sign-in. To correct or remove it, email info@ikaruscareer.com.
• Human support: a person — not a bot — answers info@ikaruscareer.com and can review any score, label, or account decision that affects you.

• We publish general, impersonal, educational commentary about macroeconomic conditions and asset valuation ranges. We do not give personal investment recommendations, and we will not do so in v1.
• We do not handle client money or assets, execute or arrange transactions, manage portfolios, or receive commission from any product provider.
• Content is not directed at any jurisdiction where its publication would require authorisation. Where local rules demand it, we will restrict access or add jurisdiction-specific notices.
• Wording and disclaimers are kept under legal review before each material feature launch; any feature that would cross into personalised advice is held until authorised or removed from scope.
• See also the Risk Disclaimer.

Our data map, retention schedule, deletion and export route, consent record, subprocessor list, model-training policy, and storage encryption are set out in the Privacy Policy.

Summary: we collect account identity, votes, and basic technical logs; we never sell personal data; we never train AI models on your personal data; documents and databases are encrypted at rest and in transit; you can request an export or deletion at any time by email.

Operator contact for governance, AI output reports, safety incidents, moderation appeals, and data requests: info@ikaruscareer.com.

Acknowledgement within 5 working days; safety incidents within 24 hours.