| CoPilot 🔗 | ChatGPT 🔗 | Claude 🔗 |
|---|
ℹ️AI Model Metadata
the nature of the AI in use- architecture
- hyperparameters
- training sources
- versioning
- performance metrics
| ℹ️AI Model Metadata
the nature of the AI in use- architecture
- hyperparameters
- training sources
- version numbers & provenance
- evaluation metrics
- usage: e.g. licenses, limitations
| ℹ️AI Model Metadata
the nature of the AI in use- architecture
- hyperparameters
- training sources
- version numbers & provenance
- performance metrics
- usage: e.g. limitations
|
ℹ️AI Data
the data used to make the AI function- provenance: origins, transforms, flows, licenses
crucial for debugging and compliance.
| ℹ️AI Datasets
the data used to make the AI function- provenance: origins, transforms, flows, licenses
- descriptive: dataset name, creator, creation date, format, size.
- quality: labeling accuracy, bias detection results, completeness.
- ethical: consent info, privacy considerations, restrictions
| ℹ️AI Training Metadata
the data used to make the AI function- provenance: origins, transforms, flows, licenses
- descriptive: training hyperparameters and configs
- quantitative: compute resources, training duration, costs, testing results
|
ℹ️AI Inference Metadata
generated during AI usage- usage: when, where, what, how ai made/changed content
- metrics: timestamps, resources, models used, confidence scores
- provenance: sources
| ℹ️AI Generated Content
generated during AI usage- usage: when, where, what, how ai made/changed content
- metrics: timestamps, resources, models used
- provenance: sources, ai-generated tagging, watermarking
- attribution: model, services & sources backlinks
| ℹ️AI Inference Metadata
generated during AI usage- usage: when, where, what, how ai made/changed content
- metrics: timestamps, resources, models used
- provenance: sources, ai-generated tagging, watermarking
|
ℹ️AI Feature Metadata
a copilot category- feature types
- encoding strategies
- statistical properties & relationships
| ℹ️AI Governance & Compliance
a chatGPT category- Transparency Auditing AI
- Reproducibility
- Ethical compliance
| ℹ️AI Content Metadata
a Claude category- Provenance markers
- Generation prompts & params
- QA & human review status
- Licensing and usage
|