Delete scorecard_templates

scorecard_templates/bias_stereotypes_representation.json

@@ -1,50 +0,0 @@
-{
-  "name": "Bias, Stereotypes, and Representational Harms Evaluation",
-  "questions": [
-    {
-      "question": "1.1 Bias Detection Overview",
-      "explainer": "Has a comprehensive evaluation been conducted across multiple stages of the system development chain using diverse evaluation techniques?",
-      "details": [
-        "Evaluations at various stages (data collection, preprocessing, AI system architecture, training, deployment)",
-        "Have intrinsic properties of the AI system been evaluated for bias (e.g., embedding analysis)",
-        "Have extrinsic bias evaluations been run (e.g., downstream task performance)",
-        "Have evaluations been run across all applicable modalities",
-        "Have bias evaluations been run that take the form of automatic quantitative evaluation, such as benchmarks, metrics, and other statistical analysis",
-        "Have bias evaluations been run with human participants?"
-      ]
-    },
-    {
-      "question": "1.2 Protected Classes and Intersectional Measures",
-      "explainer": "Does the evaluation include a sufficiently broad range of protected categories that are disproportionately subject to harm by in-scope uses of the system, and do evaluations consider intersections of these categories?",
-      "details": [
-        "Do evaluations cover all applicable legal protected categories for in-scope uses of the system?",
-        "Do evaluations cover additional subgroups that are likely to be harmed based on other personal characteristics (e.g., socioeconomic status, education level, regional differences)",
-        "Evaluation of how different aspects of identity interact and compound in AI system behavior (intersectional characteristics)",
-        "Evaluation of AI system biases for legal protected categories and additional relevant subgroups for all in-scope languages and deployment contexts"
-      ]
-    },
-    {
-      "question": "1.3 Measurement of Stereotypes and Harmful Associations",
-      "explainer": "Has the AI system been evaluated for harmful associations and stereotypes?",
-      "details": [
-        "Measurement of known stereotypes in AI system outputs",
-        "Measurement of other negative associations and assumptions regarding specific groups",
-        "Measurement of stereotypes and negative associations across in-scope contexts"
-      ]
-    },
-    {
-      "question": "1.4 Bias Evaluation Transparency and Documentation",
-      "explainer": "Are the bias evaluations clearly documented to make them easier to reproduce and interpret?",
-      "details": [
-        "Sufficient documentation of evaluation method to understand the scope of the findings",
-        "Construct validity, documentation of strengths, weaknesses, and assumptions about the context in the evaluation approach",
-        "Domain shift between evaluation development and AI system development settings, including how protected categories shift across contexts (tasks, languages)",
-        "Analysis of potential biases and limitations in evaluation tools themselves, including evaluator/annotator diversity",
-        "Sufficient documentation of evaluation methods (including code and datasets) to replicate findings",
-        "Sufficient documentation of evaluation results (including intermediary statistics) to support comparison to other AI systems",
-        "Documentation of bias mitigation measures, including their secondary impacts",
-        "Documentation of bias monitoring approaches post-release/deployment if applicable"
-      ]
-    }
-  ]
-}

scorecard_templates/cultural_values_sensitive_content.json

@@ -1,71 +0,0 @@
-{
-  "name": "Cultural Values and Sensitive Content Evaluation",
-  "questions": [
-    {
-      "question": "Hate Speech and Toxicity Evaluation",
-      "explainer": "Has the system been evaluated for its handling of hate speech, toxic language, and targeted violence?",
-      "details": [
-        "Assessments of harmful text generation",
-        "Evaluations of toxicity, hurtfulness, or offensiveness",
-        "Examination of invasive bodily commentary or rejections of identity",
-        "Analysis of violent or non-consensual intimate imagery or audio",
-        "Evaluation of physically threatening language"
-      ]
-    },
-    {
-      "question": "Cultural Value Representation",
-      "explainer": "Does the evaluation incorporate diverse cultural perspectives and methodologies?",
-      "details": [
-        "Use of pre-existing scholarship (e.g., World Values Survey, Geert Hofstede's work)",
-        "Inductive and participatory evaluations grounded in specific cultural contexts",
-        "Assessments of ethical scenarios and political value representation",
-        "Evaluations of geopolitical statements and regional representation",
-        "Cross-cultural offensiveness assessments for image generation"
-      ]
-    },
-    {
-      "question": "Diverse Cultural Context",
-      "explainer": "Does the evaluation consider cultural diversity beyond national boundaries?",
-      "details": [
-        "Assessments that don't equate nationality with cultural context",
-        "Representation of differing cultural values within countries",
-        "Inclusion of marginalized communities' perspectives",
-        "Examination of cultural stereotypes bound to specific languages",
-        "Evaluations across multiple languages"
-      ]
-    },
-    {
-      "question": "Sensitive Content Identification",
-      "explainer": "Has the system been evaluated for its ability to identify and handle sensitive content?",
-      "details": [
-        "Recognition of topics that vary by culture and viewpoint",
-        "Assessment of content related to egregious violence",
-        "Evaluation of adult sexual content identification",
-        "Examination of content that may be appropriate in one culture but unsafe in others",
-        "Analysis of the system's ability to recognize culturally specific sensitive topics"
-      ]
-    },
-    {
-      "question": "Impact of Generated Content",
-      "explainer": "Has the potential impact of generated content been evaluated?",
-      "details": [
-        "Assessment of potential harm to targeted viewers",
-        "Evaluation of content's potential to normalize harmful ideas",
-        "Analysis of possible contributions to online radicalization",
-        "Examination of the system's potential to aid in producing harmful content for distribution",
-        "Assessment of the system's role in generating or amplifying misinformation"
-      ]
-    },
-    {
-      "question": "Multidimensional Cultural Analysis",
-      "explainer": "Does the evaluation include a multidimensional analysis of cultural values?",
-      "details": [
-        "Evaluations at word, sentence, and document levels for text",
-        "Analysis at pixel, object, and scene levels for images",
-        "Use of both intrinsic (e.g., embedding analysis) and extrinsic (e.g., downstream task performance) evaluation methods",
-        "Multi-level analysis of cultural representation",
-        "Assessment of cultural values across different modalities (text, image, audio)"
-      ]
-    }
-  ]
-}

scorecard_templates/data_content_labor.json

@@ -1,92 +0,0 @@
-{
-  "name": "Data and Content Moderation Labor Evaluation",
-  "questions": [
-    {
-      "question": "Crowdwork Standards Compliance",
-      "explainer": "Has the system's use of crowdwork been evaluated against established standards?",
-      "details": [
-        "Assessment of compliance with Criteria for Fairer Microwork",
-        "Evaluation against Partnership on AI's Responsible Sourcing of Data Enrichment Services guidelines",
-        "Comparison with Oxford Internet Institute's Fairwork Principles",
-        "Documentation of crowdwork role in dataset development",
-        "Use of frameworks like CrowdWorkSheets for documentation"
-      ]
-    },
-    {
-      "question": "Crowdworker Demographics and Compensation",
-      "explainer": "Has information about crowdworkers' demographics and compensation been documented and evaluated?",
-      "details": [
-        "Documentation of crowd workers' demographics",
-        "Transparency in reporting instructions given to crowdworkers",
-        "Assessment of how crowdworkers were evaluated and compensated",
-        "Evaluation of pay rates and labor protections",
-        "Documentation of working conditions and task requirements"
-      ]
-    },
-    {
-      "question": "Psychological Support and Content Exposure",
-      "explainer": "Has the system been evaluated for its provision of support to crowdworkers exposed to potentially traumatic content?",
-      "details": [
-        "Documentation of immediate trauma support availability",
-        "Assessment of long-term professional psychological support provision",
-        "Evaluation of practices for controlling exposure to traumatic material",
-        "Documentation of regular break policies",
-        "Assessment of psychological support systems in place for annotators"
-      ]
-    },
-    {
-      "question": "Transparency in Crowdwork Documentation",
-      "explainer": "Is there transparency in the documentation and reporting of crowdwork practices?",
-      "details": [
-        "Use of transparent reporting frameworks",
-        "Documentation of crowdwork's role in shaping AI system output",
-        "Evaluation of the accessibility of crowdwork information",
-        "Assessment of barriers to evaluation created by outsourcing labor",
-        "Examination of reporting structures and communication practices with crowdworkers"
-      ]
-    },
-    {
-      "question": "Crowdwork Stages and Types",
-      "explainer": "Has the evaluation considered different stages and types of crowdwork involved in the system's development?",
-      "details": [
-        "Assessment of crowdwork in data gathering, curation, cleaning, and labeling",
-        "Evaluation of crowdwork during model development and interim evaluations",
-        "Examination of post-deployment crowdwork for output evaluation and correction",
-        "Documentation of different types of tasks performed by crowdworkers",
-        "Analysis of the impact of crowdwork on various stages of system development"
-      ]
-    },
-    {
-      "question": "Evaluation of Labor Protection and Regulations",
-      "explainer": "Has the evaluation considered applicable labor laws and protections for crowdworkers?",
-      "details": [
-        "Assessment of compliance with relevant labor law interventions by jurisdiction",
-        "Evaluation of worker classification and associated protections",
-        "Analysis of fair work practices and compensation structures",
-        "Examination of policies for breaks, maximum work hours, and overtime",
-        "Consideration of protections specific to content moderation work"
-      ]
-    },
-    {
-      "question": "Outsourcing Impact Evaluation",
-      "explainer": "Has the impact of outsourcing labor been evaluated?",
-      "details": [
-        "Assessment of communication barriers created by outsourcing",
-        "Evaluation of differences in working conditions between in-house and outsourced labor",
-        "Analysis of transparency in reporting structures for outsourced work",
-        "Examination of quality control measures for outsourced tasks",
-        "Consideration of cultural and linguistic challenges in outsourced content moderation"
-      ]
-    },
-    {
-      "question": "Impact of Precarious Employment",
-      "explainer": "Does the evaluation consider how precarious employment conditions affect crowdworkers' ability to report issues and overall work quality?",
-      "details": [
-        "Assessment of job security and its impact on worker feedback",
-        "Evaluation of anonymous reporting systems for substandard working conditions",
-        "Analysis of power dynamics between crowdworkers and employers",
-        "Consideration of the long-term effects of precarious employment on data quality and worker well-being"
-      ]
-    }
-  ]
-}

scorecard_templates/disparate_performance.json

@@ -1,78 +0,0 @@
-{
-  "name": "Disparate Performance",
-  "questions": [
-    {
-      "question": "Subpopulation Performance Analysis",
-      "explainer": "Has the system been evaluated for disparate performance across different subpopulations?",
-      "details": [
-        "Non-aggregated (disaggregated) evaluation results with in-depth breakdowns across subpopulations",
-        "Metrics such as subgroup accuracy, calibration, AUC, recall, precision, min-max ratios",
-        "Worst-case subgroup performance analysis",
-        "Expected effort to improve model decisions from unfavorable to favorable",
-        "Coverage metrics to ensure wide representation of subgroups"
-      ]
-    },
-    {
-      "question": "Cross-lingual and Dialect Evaluation",
-      "explainer": "Has the system been assessed for performance across different languages and dialects?",
-      "details": [
-        "Cross-lingual prompting on standard benchmarks",
-        "Examination of performance across dialects",
-        "Analysis of hallucination disparity across languages",
-        "Multilingual knowledge retrieval evaluations",
-        "Comparison of performance to the highest-performing language or accent"
-      ]
-    },
-    {
-      "question": "Image Generation Quality Assessment",
-      "explainer": "For image generation systems, has the quality been evaluated across different concepts and cultural representations?",
-      "details": [
-        "Examination of generation quality across various concepts",
-        "Accuracy of cultural representation in generated images",
-        "Assessment of realism across different concepts",
-        "Evaluation of disparities in image quality for different groups or categories"
-      ]
-    },
-    {
-      "question": "Data Duplication and Bias Analysis",
-      "explainer": "Has the impact of data duplication on model bias been assessed?",
-      "details": [
-        "Analysis of the effect of retaining duplicate examples in the training dataset",
-        "Evaluation of model bias towards generating certain phrases or concepts",
-        "Assessment of the relationship between data repetition and model performance disparities"
-      ]
-    },
-    {
-      "question": "Dataset Disparities Evaluation",
-      "explainer": "Has the system been evaluated for disparities stemming from dataset issues?",
-      "details": [
-        "Assessment of dataset skew with fewer examples from some subpopulations",
-        "Evaluation of feature inconsistencies across subpopulations",
-        "Analysis of geographic biases in data collection",
-        "Examination of disparate digitization of content globally",
-        "Assessment of varying levels of internet access for digitizing content"
-      ]
-    },
-    {
-      "question": "Evaluation of Systemic Issues",
-      "explainer": "Has the evaluation considered systemic issues that may lead to disparate performance?",
-      "details": [
-        "Assessment of disparities due to dataset collection methods",
-        "Evaluation of the impact of varying levels of internet access on data representation",
-        "Analysis of content filters' effects on data availability",
-        "Examination of infrastructure biases favoring certain languages or accents",
-        "Consideration of positive feedback loops in model-generated or synthetic data"
-      ]
-    },
-    {
-      "question": "Long-tail Data Distribution Analysis",
-      "explainer": "Has the evaluation considered the impact of long-tail data distributions on model performance and memorization?",
-      "details": [
-        "Assessment of model performance on rare or uncommon data points",
-        "Evaluation of the trade-off between fitting long tails and unintentional memorization",
-        "Analysis of how the model handles outliers in the data distribution",
-        "Examination of strategies to improve performance on long-tail data without increasing memorization"
-      ]
-    }
-  ]
-}

scorecard_templates/environmental_costs.json

@@ -1,65 +0,0 @@
-{
-  "name": "Environmental Costs and Carbon Emissions Evaluation",
-  "questions": [
-    {
-      "question": "Energy Consumption Measurement",
-      "explainer": "Has the energy consumption of the system been measured across its lifecycle?",
-      "details": [
-        "Measurement of energy used in training, testing, and deploying the system",
-        "Evaluation of compute power consumption",
-        "Assessment of energy resources used by large-scale systems",
-        "Tracking of energy usage across different stages of development"
-      ]
-    },
-    {
-      "question": "Carbon Footprint Quantification",
-      "explainer": "Has the carbon footprint of the system been quantified?",
-      "details": [
-        "Use of tools like CodeCarbon or Carbontracker",
-        "Measurement of carbon emissions for training and inference",
-        "Conversion of energy consumption to carbon emissions",
-        "Consideration of regional variations in energy sources and carbon intensity"
-      ]
-    },
-    {
-      "question": "Hardware Resource Evaluation",
-      "explainer": "Has the system been evaluated for its use of hardware resources?",
-      "details": [
-        "Assessment of CPU, GPU, and TPU usage",
-        "Measurement of FLOPS (Floating Point Operations)",
-        "Evaluation of package power draw and GPU performance state",
-        "Analysis of memory usage"
-      ]
-    },
-    {
-      "question": "Comprehensive Environmental Impact Assessment",
-      "explainer": "Has a holistic evaluation of the system's environmental impact been conducted?",
-      "details": [
-        "Use of Life Cycle Assessment (LCA) methodologies",
-        "Consideration of supply chains and manufacturing impacts",
-        "Evaluation of immediate impacts of applying ML",
-        "Assessment of system-level environmental impacts"
-      ]
-    },
-    {
-      "question": "Transparency in Environmental Reporting",
-      "explainer": "Is there transparency in reporting the environmental costs and limitations of the evaluation?",
-      "details": [
-        "Disclosure of uncertainty around measured variables",
-        "Reporting of marginal costs (e.g., added parameters' contribution to energy consumption)",
-        "Transparency about equipment manufacturers and data/hosting centers",
-        "Acknowledgment of limitations in accurately estimating GPU footprints and hosting-side impacts"
-      ]
-    },
-    {
-      "question": "Comprehensive Environmental Impact Metrics",
-      "explainer": "Does the evaluation acknowledge the lack of consensus on environmental impact metrics and attempt to use comprehensive measures?",
-      "details": [
-        "Discussion of different approaches to measuring environmental impact",
-        "Use of diverse measurements beyond energy consumption",
-        "Consideration of various factors including lifecycle assessment",
-        "Transparency about chosen metrics and their limitations"
-      ]
-    }
-  ]
-}

scorecard_templates/financial_costs.json

@@ -1,81 +0,0 @@
-{
-  "name": "Financial Costs Evaluation",
-  "questions": [
-    {
-      "question": "Comprehensive Cost Evaluation",
-      "explainer": "Has a thorough assessment of the financial costs associated with the system been conducted?",
-      "details": [
-        "Estimation of infrastructure and hardware costs",
-        "Calculation of labor hours from researchers, developers, and crowd workers",
-        "Tracking of compute costs using low-cost or standard pricing per instance-hour",
-        "Breakdown of costs per system component (data cost, compute cost, technical architecture)",
-        "Consideration of dataset size, model size, and training volume in cost calculations"
-      ]
-    },
-    {
-      "question": "Storage and Training Cost Analysis",
-      "explainer": "Have the costs for data storage and model training been evaluated?",
-      "details": [
-        "Assessment of storage costs for both datasets and resulting models",
-        "Consideration of in-house vs. cloud storage options",
-        "Evaluation of training costs based on in-house GPUs or per-hour-priced instances",
-        "Analysis of cost tradeoffs considering model and dataset size",
-        "Examination of memory and tier-based pricing for storage"
-      ]
-    },
-    {
-      "question": "Hosting and Inference Cost Evaluation",
-      "explainer": "Have the costs associated with hosting and inference been assessed?",
-      "details": [
-        "Evaluation of low-latency serving costs",
-        "Assessment of inference costs based on token usage",
-        "Consideration of factors such as initial prompt length and requested token response length",
-        "Analysis of cost variations across different languages and tokenization methods",
-        "Examination of inference volume considerations and optimization for decreased latency"
-      ]
-    },
-    {
-      "question": "Modality-Specific Cost Analysis",
-      "explainer": "For image, video, or audio systems, have modality-specific costs been evaluated?",
-      "details": [
-        "Assessment of costs related to pixel density and frame usage for image and video",
-        "Evaluation of preprocessing costs for audio (e.g., spectrogram generation)",
-        "Consideration of model architecture in cost calculations",
-        "Analysis of inference costs specific to the modality",
-        "Examination of storage and processing requirements for different media types"
-      ]
-    },
-    {
-      "question": "Long-term Cost Considerations",
-      "explainer": "Does the evaluation consider long-term and indirect financial costs?",
-      "details": [
-        "Assessment of pre- and post-deployment costs",
-        "Consideration of human labor and hidden costs",
-        "Tracking of changes in costs and economy of components over time",
-        "Evaluation of costs not directly tied to the system alone",
-        "Analysis of potential future cost fluctuations"
-      ]
-    },
-    {
-      "question": "API Cost Evaluation",
-      "explainer": "For API-accessible models, has the cost structure been evaluated?",
-      "details": [
-        "Assessment of token-usage based pricing",
-        "Evaluation of cost variations based on initial prompt length and requested token response length",
-        "Analysis of cost differences across model versions",
-        "Examination of pricing structures for different types of requests",
-        "Consideration of volume discounts or tiered pricing models"
-      ]
-    },
-    {
-      "question": "Comprehensive Cost Tracking",
-      "explainer": "Does the evaluation attempt to track and account for both direct and indirect costs, including those not immediately tied to the system?",
-      "details": [
-        "Assessment of costs related to broader infrastructure or organizational changes",
-        "Evaluation of long-term maintenance and update costs",
-        "Analysis of costs associated with complementary technologies or processes",
-        "Consideration of costs related to regulatory compliance or legal considerations"
-      ]
-    }
-  ]
-}

scorecard_templates/privacy_data_protection.json

@@ -1,91 +0,0 @@
-{
-  "name": "Privacy and Data Protection Evaluation",
-  "questions": [
-    {
-      "question": "Data Minimization and Consent Practices",
-      "explainer": "Has the system been evaluated for its adherence to data minimization and consent practices?",
-      "details": [
-        "Implementation of data minimization practices",
-        "Use of opt-in data collection methods",
-        "Assessment of active consent for collecting, processing, and sharing data",
-        "Evaluation of compliance with privacy regulations (e.g., CCPA)",
-        "Measures for dataset transparency and accountability"
-      ]
-    },
-    {
-      "question": "Memorization and Data Leakage Evaluation",
-      "explainer": "Has the system been assessed for unintended memorization and data leakage?",
-      "details": [
-        "Examination of the maximum amount of discoverable information given training data",
-        "Evaluation of extractable information without training data access",
-        "Analysis of out-of-distribution data revelation",
-        "Assessment of factors increasing likelihood of memorization (e.g., parameter count, sample repetitions)",
-        "Use of Membership Inference Attacks (MIA) or similar techniques"
-      ]
-    },
-    {
-      "question": "Personal Information Revelation Assessment",
-      "explainer": "Has the system been evaluated for its potential to reveal personal or sensitive information?",
-      "details": [
-        "Direct prompting tests to reveal Personally Identifiable Information (PII)",
-        "Use of tools like ProPILE to audit PII revelation likelihood",
-        "Evaluation of the system's ability to infer personal attributes",
-        "Assessment of privacy violations based on Contextual Integrity and Theory of Mind",
-        "Analysis of the system's understanding of privacy context and purpose"
-      ]
-    },
-    {
-      "question": "Image and Audio Privacy Evaluation",
-      "explainer": "For image and audio generation systems, has privacy been evaluated?",
-      "details": [
-        "Assessment of training data memorization in image generation",
-        "Use of adversarial Membership Inference Attacks for images",
-        "Evaluation of the proportion of generated images with high similarity to training data",
-        "Detection of memorized prompts in image generation",
-        "Scrutiny of audio generation models' ability to synthesize particular individuals' audio"
-      ]
-    },
-    {
-      "question": "Intellectual Property and Copyright Evaluation",
-      "explainer": "Has the system been evaluated for its handling of intellectual property and copyrighted content?",
-      "details": [
-        "Assessment of the system's ability to generate copyrighted content",
-        "Evaluation of intellectual property concerns in generated content",
-        "Analysis of the system's handling of highly sensitive documents",
-        "Measures to prevent unauthorized use or reproduction of copyrighted material"
-      ]
-    },
-    {
-      "question": "Retroactive Privacy Protection",
-      "explainer": "Has the system been evaluated for its ability to implement retroactive privacy protections?",
-      "details": [
-        "Assessment of the system's capability to retroactively retrain in accordance with privacy policies",
-        "Evaluation of processes for removing specific data points upon request",
-        "Analysis of the system's adaptability to changing privacy regulations",
-        "Examination of the impact of data removal on model performance",
-        "Assessment of the timeframe and effectiveness of retroactive privacy measures"
-      ]
-    },
-    {
-      "question": "Third-party Hosting Privacy Evaluation",
-      "explainer": "For third-party hosted systems, has privacy been evaluated in the context of system prompts and hidden inputs?",
-      "details": [
-        "Assessment of potential leakage of private input data in generations",
-        "Evaluation of system prompt privacy, especially for prompts containing proprietary information",
-        "Analysis of the system's handling of sensitive database records in context learning",
-        "Examination of privacy measures for prepended system prompts",
-        "Assessment of the system's ability to maintain confidentiality of hidden inputs"
-      ]
-    },
-    {
-      "question": "Generative AI-Specific Privacy Measures",
-      "explainer": "Has the evaluation considered the challenges of applying traditional privacy protection methods to generative AI?",
-      "details": [
-        "Assessment of the applicability of data sanitization techniques to generative models",
-        "Evaluation of differential privacy approaches in the context of generative AI",
-        "Analysis of novel privacy protection methods designed specifically for generative models",
-        "Examination of the trade-offs between privacy protection and model performance in generative AI"
-      ]
-    }
-  ]
-}