Last Updated: 2026-04-12
This document describes dbl9 GmbH's ("dbl9") approach to trust, safety, and responsible AI within the dbl9 Platform platform. It supplements the dbl9 Terms of Service, Acceptable Use Policy, and Privacy Policy.
1. OUR COMMITMENT
dbl9 is committed to developing and deploying AI technology in a manner that is safe, transparent, and respectful of user rights. We design our AI Features with the following principles:
- Safety: AI Features are designed to minimize the risk of harmful outputs and to include appropriate safeguards.
- Transparency: We are transparent about how AI Features work, their limitations, and how Customer Data is used.
- Privacy: Customer Data is processed in accordance with our Privacy Policy and Data Processing Addendum. We do not train general-purpose AI models on Customer Data without explicit consent.
- Fairness: We take steps to identify and mitigate bias in AI systems.
- Accountability: We maintain human oversight over AI systems and provide mechanisms for users to report issues.
2. HOW AI FEATURES WORK
2.1. AI Models
dbl9 Platform uses third-party and/or proprietary AI models to power its AI Features. These models may include:
- Large Language Models (LLMs), natural language processing models, and custom fine-tuned models
2.2. Data Processing
- Input Data: Data submitted to AI Features is processed to generate output and is not retained beyond the duration necessary to deliver the response, unless Customer configures retention settings.
- No Training Without Consent: dbl9 does not use Customer Data or Input Data to train, fine-tune, or improve general-purpose AI models without Customer's prior written consent.
- Aggregated Data: dbl9 may use aggregated, anonymized data to monitor system performance, improve safety measures, and enhance the Service, provided such data cannot reasonably identify any individual or Customer.
2.3. Third-Party AI Providers
The Service may use AI models provided by third-party providers. These providers are listed on our Sub-Processors page and are contractually bound to equivalent data protection standards.
3. SAFETY MEASURES
3.1. Content Filtering
- AI Features include content safety filters designed to prevent the generation of harmful, illegal, or inappropriate content.
- Content safety filters screen both input and output for prohibited content categories as defined in the Acceptable Use Policy.
3.2. Rate Limiting and Abuse Prevention
- Technical controls are in place to detect and prevent abuse of AI Features, including rate limiting, anomaly detection, and automated monitoring.
3.3. Human Oversight
- dbl9 maintains human oversight processes for AI systems, including monitoring for systematic issues and review of flagged content.
- Customers are encouraged to implement their own human review processes for AI-generated output, particularly for high-stakes use cases.
4. LIMITATIONS AND DISCLAIMERS
4.1. AI Features are designed to assist users but are not a substitute for human judgment, expertise, or professional advice.
4.2. AI Output may contain errors, inaccuracies, or biases. Users should independently verify AI Output before relying on it.
4.3. AI Features may not perform equally well across all languages, domains, or use cases.
4.4. dbl9 continuously works to improve the safety and accuracy of AI Features, but cannot guarantee that all harmful or inaccurate outputs will be prevented.
5. COMPLIANCE WITH AI REGULATIONS
dbl9 monitors and works to comply with applicable AI regulations, including:
- EU AI Act: dbl9 monitors developments under the EU AI Act and will implement compliance measures as requirements become applicable.
- NIST AI Risk Management Framework: dbl9 references the NIST AI Risk Management Framework in its AI governance processes.
- Other Regulations:
6. REPORTING CONCERNS
If you encounter harmful, inappropriate, or concerning AI-generated content, or if you have concerns about the safety or fairness of our AI Features, please report them to:
Email: support@dbl99.de
We review all reports and take appropriate action in accordance with our policies.
7. FREQUENTLY ASKED QUESTIONS
Q: Does dbl9 use my data to train AI models? A: No. We do not use Customer Data or Input Data to train general-purpose AI models without your prior written consent. See our Privacy Policy and DPA for details.
Q: Can I opt out of AI Features? A: Yes. Customers can disable AI Features in their Account settings. Contact support@dbl99.de for assistance.
Q: How does dbl9 address bias in AI? A: dbl9 evaluates AI models for bias during development and deployment, and implements mitigation measures where bias is identified.
Q: Who is responsible for AI Output? A: Customers are responsible for reviewing and validating AI Output before use. dbl9 provides the tools; Customers are responsible for how they use the output. See our Terms of Service, Section 3.
dbl9 AI Trust and Safety v1.0