AdPilot AI is committed to building advertising tools that are useful, safe and respectful of users and the people depicted in ad content. This Responsible AI Policy describes the principles and practical limits that guide how we design, deploy and maintain AI features. It complements our AI Usage Policy and Acceptable Use Policy.
1. Limitations of AI
- Hallucinations: generative models can fabricate facts, names, statistics, brand attributes and product details. Treat AI outputs as drafts, not as verified facts.
- Bias: models can reflect biases present in their training data. Outputs may underrepresent or misrepresent demographic groups, languages or cultures.
- Drift: model behavior may change over time as providers update their models. Outputs for the same prompt may differ between runs.
- No real-time knowledge: models may not have current knowledge of events, prices, regulations or market conditions.
2. Human review
Every output should be reviewed by a human before publication. AdPilot is a tool to assist marketers; it does not replace the editorial, legal or compliance judgment that advertisers must apply to their campaigns.
3. Transparency
- We disclose which model providers we use and route requests through (see our Privacy Policy).
- We label AI-generated artifacts in the dashboard, including the model name and generation timestamp where practical.
- Where applicable law or platform rules require disclosure of AI involvement in advertising, you are responsible for making that disclosure.
4. Safety features
- We apply prompt-level guardrails to discourage clearly harmful generations.
- We rely on, and inherit, the safety filters of our model providers.
- We enforce per-user and per-IP rate limits to reduce abuse and accidental over-generation.
- We log generation events for audit and abuse-investigation purposes, in line with our retention policy.
5. User responsibility
- You are responsible for the prompts and assets you submit and the outputs you publish.
- You must not use the Service to generate content prohibited by our Acceptable Use Policy, including non-consensual deepfakes, defamatory content, or misleading advertising.
- You must obtain all necessary rights and consents for any individual or third-party material depicted in your inputs.
6. Continuous improvement
We monitor model performance, user feedback and reported incidents, and update our prompts, guardrails and provider selection as needed. We welcome reports of harmful or biased outputs at abuse@adpilot.ai.
7. No professional advice
AI outputs do not constitute legal, financial, medical or other professional advice. Do not rely on them as a substitute for advice from a qualified professional.