AI Usage & Responsible AI Policy

CareerOS-AI uses large language models and other AI systems in the following features:

• Resume parsing, scoring, and rewriting.
• Cover-letter generation.
• Chat with your AI career coach.
• Career roadmap and insight generation.
• Salary benchmarking commentary (numbers come from our data; interpretation from AI).
• Interview question generation and answer evaluation.
• Salary-negotiation script writing.
• Financial-plan recommendations and commentary.
• Skills and personality assessment interpretation.
• Support-agent responses in the in-app help widget.

We currently rely on external model providers, including Anthropic (Claude), OpenAI (GPT), and Google (Gemini). We select models per feature based on quality, latency, cost, and regional availability.

Our provider agreements require:

• No use of your prompts or CareerOS content to train the provider's public models.
• Short retention windows (typically 30 days or less) for abuse-monitoring only.
• Regional data-residency controls where offered.

Every user-facing AI prompt automatically receives a localized context block built from your profile — country, region, city, currency, timezone, tax authority, retirement vehicles, employment framework, spelling, and date format. This lets the AI give locally-accurate advice instead of US defaults.

We do not include other users' data, private business rules, or your billing details in prompts.

AI systems make mistakes. We design CareerOS-AI to reduce them but we do not eliminate them. Concretely:

• Numbers that change year-to-year (tax rates, contribution limits, minimum wage, retirement age, filing deadlines) are treated as estimates and should be verified with the appropriate authority.
• AI may generate confident-sounding text that is wrong. Always sanity-check important claims.
• AI is not a substitute for a licensed CPA, tax preparer, financial advisor, lawyer, immigration consultant, or licensed mental-health professional.
• We provide educational content, not regulated advice.

• You choose what to save. AI outputs are shown to you before we persist them.
• You can regenerate any AI output.
• You can delete AI-generated content from your account at any time.
• You can export your data from Settings.

You agree not to use CareerOS-AI to:

• Fabricate credentials, employment history, or achievements you do not hold.
• Impersonate another person in resumes, portfolios, or communications.
• Generate content that discriminates, harasses, or targets protected classes.
• Circumvent hiring, licensing, or educational integrity requirements.
• Attempt to reverse-engineer or systematically extract our prompts, tools, or datasets.
• Train competing AI systems on CareerOS content or outputs.

Violations are subject to account suspension per our Acceptable Use Policy.

CareerOS-AI does not make consequential automated decisions about you (for example, we do not deny access, terminate accounts, or refuse refunds solely because an AI said so). Consequential decisions are reviewed by a human on our team.

LLMs inherit biases from their training data. We take steps to reduce user-visible harm:

• Localization to avoid US-centric defaults.
• Prompt-level rules against demographic stereotyping in salary, career, or hiring recommendations.
• Periodic red-team review of outputs.

If you spot an unfair or biased output, report it via the in-app support widget — we take these reports seriously.

We do not use your account data, prompts, resumes, chats, or financial inputs to train third-party foundation models.

We may use aggregated, de-identified metrics (feature usage, latency, error rates) to improve the product. Where we manually review a specific case, it is only for support, safety, or debugging — and only by authorized staff under confidentiality.

Some CareerOS-AI features touch regulated areas:

• Financial planning — educational only; not investment or tax advice.
• Immigration — informational only; consult a licensed immigration lawyer or RCIC.
• Employment law — informational only; consult a licensed employment lawyer.
• Mental-health topics — CareerOS is not a therapy service. If you are in crisis, contact your local emergency number or a crisis line.

Please report anything that looks like an AI-generated harm — hallucinated facts about a person, biased outputs, unsafe advice, or a jailbreak — to support@careeros-ai.com. Include the feature, timestamp, and (if possible) a copy of the output. We treat these reports as high priority.

We will update this policy as models, providers, and best practices evolve. Material changes are communicated on the website. The "Last Updated" date reflects the current version.