AI Governance & Use Policy

1. Innledning

AhaSlides provides AI-powered features to help users generate slides, enhance content, group responses, and more. This AI Governance & Use Policy outlines our approach to responsible AI usage, including data ownership, ethical principles, transparency, support, and user control.

2. Ownership and Data Handling

• User Ownership: All user-generated content, including content created with the help of AI features, belongs solely to the user.
• AhaSlides IP: AhaSlides retains all rights to its logo, brand assets, templates, and platform-generated interface elements.
• Databehandling:
  • AI features may send inputs to third-party model providers (e.g., OpenAI) for processing. Data is not used to train third-party models unless explicitly stated and consented to.
  • Most AI features do not require personal data unless it is intentionally included by the user. All processing is done in accordance with our Privacy Policy and GDPR commitments.
• Exit and Portability: Users may export slide content or delete their data at any time. We do not currently offer automated migration to other providers.

3. Bias, Fairness, and Ethics

• Bias Mitigation: AI models may reflect biases present in training data. While AhaSlides uses moderation to reduce inappropriate results, we do not directly control or retrain third-party models.
• Fairness: AhaSlides proactively monitors AI models to minimize bias and discrimination. Fairness, inclusivity, and transparency are core design principles.
• Ethical Alignment: AhaSlides supports responsible AI principles and aligns with industry best practices but does not formally certify to any specific regulatory AI ethics framework.

4. Åpenhet og forklaring

• Decision Process: AI-powered suggestions are generated by large language models based on context and user input. These outputs are probabilistic and not deterministic.
• User Review Required: Users are expected to review and validate all AI-generated content. AhaSlides does not guarantee accuracy or appropriateness.

5. AI System Management

• Post-Deployment Testing and Validation: A/B testing, human-in-the-loop validation, output consistency checks, and regression testing are employed to verify AI system behaviour.
• Ytelsesberegninger:
  • Accuracy or coherence (where applicable)
  • User acceptance or usage rates
  • Latency and availability
  • Complaint or error report volume
• Monitoring and Feedback: Logging and dashboards track model output patterns, user interaction rates, and flagged anomalies. Users can report inaccurate or inappropriate AI output via the UI or customer support.
• Change Management: All major AI system changes must be reviewed by the assigned Product Owner and tested in staging before production deployment.

6. User Controls and Consent

• User Consent: Users are informed when using AI features and may choose not to use them.
• Moderation: Prompts and outputs may be automatically moderated to reduce harmful or abusive content.
• Manual Override Options: Users retain the ability to delete, modify, or regenerate outputs. No action is automatically enforced without user consent.
• Feedback: We encourage users to report problematic AI outputs so we can improve the experience.

7. Performance, Testing, and Audits

• TEVV (Testing, Evaluation, Verification & Validation) tasks are carried out.
  • At every major update or retraining
  • Monthly for performance monitoring
  • Immediately upon incident or critical feedback
• Reliability: AI features are dependent on third-party services, which may introduce latency or occasional inaccuracy.

8. Integrasjon og skalerbarhet

• Scalability: AhaSlides uses scalable, cloud-based infrastructure (e.g., OpenAI APIs, AWS) to support AI features.
• Integration: AI features are embedded into the AhaSlides product interface and are not currently available via public API.

9. Støtte og vedlikehold

• Support: Users can contact hi@ahaslides.com for issues related to AI-powered features.
• Maintenance: AhaSlides may update AI features as improvements become available through providers.

10. Liability, Warranty, and Insurance

• Disclaimer: AI features are provided "as-is." AhaSlides disclaims all warranties, express or implied, including any warranty of accuracy, fitness for a particular purpose, or non-infringement.
• Limitation of Warranty: AhaSlides is not responsible for any content generated by AI features or any damages, risks, or losses — direct or indirect — resulting from reliance on AI-generated outputs.
• Insurance: AhaSlides does not currently maintain specific insurance coverage for AI-related incidents.

11. Incident Response for AI Systems

• Anomaly Detection: Unexpected outputs or behaviour flagged via monitoring or user reports are treated as potential incidents.
• Incident Triage and Containment: If the issue is confirmed, rollback or restriction may be performed. Logs and screenshots are preserved.
• Root Cause Analysis: A post-incident report is produced including root cause, resolution, and updates to testing or monitoring processes.

12. Decommissioning and End-of-Life Management

• Criteria for Decommissioning: AI systems are retired if they become ineffective, introduce unacceptable risks, or are replaced by superior alternatives.
• Archiving and Deletion: Models, logs, and related metadata are archived or securely deleted per internal retention policies.

AhaSlides' AI practices are governed under this policy and further supported by our Personvernerklæring, in line with global data protection principles including GDPR.

For questions or concerns about this policy, contact us at hei@ahaslides.com.

Finn ut mer

Besøk vår AI-hjelpesenter for vanlige spørsmål, veiledninger og for å dele tilbakemeldinger om AI-funksjonene våre.

endrings~~POS=TRUNC

• July 2025: Second version of the policy issued with clarified user controls, data handling, and AI management processes.
• Februar 2025: Første versjon av siden.