Provision Registry

140 classified provisions across 299 platforms — browse, filter, and compare.

Every clause classified by type, severity, and platform. Updated as policies change.

Start Compliance free trial Track specific clauses across platforms with provision-level alerts.
Filtering: Ai automated × High × Clear all
Inflection AI · Inflection AI Privacy Policy
This clause establishes the operational scope of data use for model development, permitting the company to leverage user interactions as training material for service improvement without requiring separate user consent per interaction.
CA-P-004145 First tracked Apr 30, 2026 Last seen Apr 30, 2026 Compare across platforms →
Cash App · Cash App Privacy Policy
The provision establishes Cash App's operational authority to conduct inferential data profiling beyond the raw data collected, creating derived attributes about users' creditworthiness and behavioral patterns. This profiling supports both service optimization and risk management functions within the platform.
CA-P-000611 First tracked Apr 3, 2026 Last seen Apr 17, 2026 Compare across platforms →
Stripe · Stripe Privacy Policy
The cross-merchant fraud detection network establishes an operational infrastructure where transaction patterns are aggregated and analyzed to identify fraud schemes that span multiple merchants. This institutional arrangement allows Stripe to maintain fraud prevention services by leveraging data signals across its customer base.
CA-P-000741 First tracked Apr 3, 2026 Last seen Apr 3, 2026 Compare across platforms →
AWS Bedrock · AWS Service Terms
The provision establishes a compliance framework in which customers retain primary responsibility for output review and lawfulness rather than AWS providing pre-validated or pre-screened outputs. This allocation of responsibility affects how disputes over output-related compliance violations would be structured.
CA-P-008312 First tracked May 10, 2026 Last seen May 11, 2026 Compare across platforms →
OpenAI · GPT-4o System Card (PDF)
This risk assessment establishes the documented baseline for GPT-4o's cybersecurity capabilities and limitations as determined through OpenAI's internal evaluation framework. The Medium risk classification indicates the model's assessed capacity to provide assistance with malicious coding activities, which informs the operational scope of the model's deployment and monitoring protocols.
CA-P-000070 First tracked Apr 3, 2026 Last seen Apr 27, 2026 Compare across platforms →
Microsoft · Responsible AI Report 2025
The provision establishes operational requirements for how Microsoft structures AI training practices, including documentation and consent alignment mechanisms. These requirements define the institutional framework Microsoft applies to govern data inputs and traceability across its AI systems.
CA-P-003120 First tracked Apr 27, 2026 Last seen Apr 27, 2026 Compare across platforms →
high Ai automated
Google Gemini · Google Generative AI Prohibited Use Policy
This provision establishes the operational scope of data use for model improvement and product development. The availability of a disable option creates a user-controlled mechanism to limit ongoing conversation storage, while the short-term retention clause establishes that some data processing continues independent of user preference settings.
CA-P-003073 First tracked Apr 18, 2026 Last seen Apr 18, 2026 Compare across platforms →
Mistral AI · Mistral AI Terms of Service
The clause creates a default opt-in training authorization for certain subscription tiers while establishing carve-outs for feedback data and moderation-flagged content as separate bases for model training use, thereby defining the scope of permitted data uses across different user categories and engagement types.
CA-P-006730 First tracked May 8, 2026 Last seen May 8, 2026 Compare across platforms →
Mistral AI · Mistral AI Privacy Policy
The clause establishes a default data usage practice for model training across Mistral's free tier products, with carve-outs for commercial and enterprise offerings. This creates distinct data handling regimes based on service tier and connection type.
CA-P-007010 First tracked May 8, 2026 Last seen May 8, 2026 Compare across platforms →
OpenAI · Privacy Policy (ROW)
The clause establishes content use for model training as the operational default, creating a mechanism whereby user interactions contribute to model development unless affirmatively disabled. This structure allocates the burden of opting out to users rather than requiring affirmative consent to training use.
CA-P-009521 First tracked May 10, 2026 Last seen May 11, 2026 Compare across platforms →
eBay · eBay Privacy Notice
The clause defines the operational basis for data collection by identifying the specific touchpoints and user actions that trigger the collection of personal information within eBay's service environment. This framing establishes the procedural foundation for subsequent data handling and processing practices described elsewhere in the privacy notice.
CA-P-005708 First tracked May 7, 2026 Last seen May 7, 2026 Compare across platforms →
Microsoft · Responsible AI Report 2025
This provision defines Microsoft's regulatory compliance obligations under EU law and establishes the institutional framework through which the company will implement AI Act requirements across its product portfolio and service delivery. It affects how Microsoft structures AI system governance, documentation practices, and user-facing disclosures in EU markets.
CA-P-000035 First tracked Apr 3, 2026 Last seen Apr 10, 2026 Compare across platforms →
Luma AI · Luma AI Terms of Service
This clause establishes that free-tier inputs become subject to broad secondary use rights by the service provider, extending beyond the immediate service delivery to include public dissemination and derivative work creation. The perpetual and irrevocable nature of the license means these rights are retained even after user-generated inputs are incorporated into the service's outputs or data aggregates.
CA-P-004094 First tracked Apr 30, 2026 Last seen Apr 30, 2026 Compare across platforms →
Kling AI · Kling AI Privacy Policy
The clause establishes the functional scope of the service's generative AI features, specifying that the platform processes facial and appearance reference data as inputs to produce generated images according to user-specified parameters.
CA-P-004060 First tracked Apr 30, 2026 Last seen Apr 30, 2026 Compare across platforms →
Microsoft · Microsoft Responsible AI Standard
This statement expresses Microsoft's commitment to fairness principles in AI system design and deployment. The operational significance is limited by the absence of specific metrics, enforcement mechanisms, or remedial procedures within the clause language itself.
CA-P-002089 First tracked Apr 4, 2026 Last seen Apr 9, 2026 Compare across platforms →
Microsoft · Responsible AI Report 2025
This provision operationalizes the entity's governance framework for AI system development by establishing fairness and non-discrimination as binding commitments within product design and deployment processes. The operational significance lies in its requirement that algorithmic systems be assessed and monitored for discriminatory outcomes across protected categories.
CA-P-000031 First tracked Apr 3, 2026 Last seen Apr 10, 2026 Compare across platforms →
Stripe · Stripe Privacy Policy
The clause establishes Stripe's operational authority to conduct behavioral profiling and risk scoring across its user base as part of fraud mitigation infrastructure. This processing activity shapes how transaction data flows through Stripe's systems and informs which transactions are flagged or declined based on algorithmic assessment.
CA-P-001876 First tracked Apr 4, 2026 Last seen Apr 9, 2026 Compare across platforms →
DeepL · DeepL Terms and Conditions
The clause establishes a differentiated data handling framework between free and paid service tiers. It specifies that system improvement activities are conditioned on service tier, creating distinct operational parameters for each user class.
CA-P-005814 First tracked May 8, 2026 Last seen May 8, 2026 Compare across platforms →
Anthropic · Anthropic API Usage Policy
The clause creates mandatory procedural controls for high-risk advisory applications by requiring human professional gatekeeping and consumer-facing transparency. These requirements establish operational gates that must be implemented before service deployment in affected use cases.
CA-P-000116 First tracked Apr 3, 2026 Last seen Apr 17, 2026 Compare across platforms →
high Ai automated
Microsoft · Responsible AI
This provision articulates an accountability framework positioning human responsibility as central to the deployment of AI systems. The operational significance depends on how this principle is implemented through specific service terms, policies, or technical controls elsewhere in the documentation.
CA-P-002071 First tracked Apr 4, 2026 Last seen Apr 9, 2026 Compare across platforms →
high Ai automated
Microsoft · Responsible AI Report 2025
Human oversight requirements create governance checkpoints within AI system operations, establishing institutional accountability structures and requiring documented review processes before certain AI outputs or decisions proceed to deployment or user-facing implementation.
CA-P-000030 First tracked Apr 3, 2026 Last seen Apr 10, 2026 Compare across platforms →
DeepL · DeepL Privacy Policy
This provision establishes a data processing practice that distinguishes free-tier users from paid subscribers, creating operational conditions under which free-tier input data is designated for internal review and model training activities rather than deleted after translation.
CA-P-004493 First tracked May 7, 2026 Last seen May 7, 2026 Compare across platforms →
TikTok · TikTok Terms of Service
The irrevocable scope of this license means TikTok retains usage rights to content even if a user deletes their account or ceases platform participation. The explicit inclusion of machine learning and algorithmic development creates an institutional basis for using user content to train AI systems without ongoing user consent or compensation.
CA-P-008577 First tracked May 10, 2026 Last seen May 11, 2026 Compare across platforms →
Luma AI · Luma AI Terms of Service
The irrevocable and perpetual nature of this license establishes that Luma retains ongoing rights to inputs even after subscription termination, provided those inputs are reflected in derivative works, aggregated data, or model improvements. This structure allocates long-term intellectual property interests in materials derived from user inputs to Luma.
CA-P-004093 First tracked Apr 30, 2026 Last seen Apr 30, 2026 Compare across platforms →
Adobe · Adobe Privacy Policy
The clause establishes the operational framework for data sharing with service providers across multiple functional categories—support, analytics, and AI processing—that collectively handle recording, storage, and analysis of user interactions and content.
CA-P-001073 First tracked Apr 3, 2026 Last seen Apr 17, 2026 Compare across platforms →
high Ai automated
OpenAI · OpenAI Terms of Use
This provision establishes differentiated data practices across user tiers, requiring explicit opt-in consent for certain user categories before their content enters model training workflows, while establishing an opt-out mechanism for free users. The clause creates distinct operational pathways for content use depending on account type and user election.
CA-P-003149 First tracked Apr 27, 2026 Last seen Apr 27, 2026 Compare across platforms →
Anthropic · Claude.ai Terms of Service
The provision establishes a conditional data use framework where the opt-out right contains defined carve-outs. This structure means users cannot obtain complete exclusion from training data use through the opt-out mechanism; certain material categories remain subject to training use according to the terms' specifications.
CA-P-007114 First tracked May 9, 2026 Last seen May 9, 2026 Compare across platforms →
Anthropic · Anthropic Privacy Policy
The provision establishes a data usage framework where model training on user-generated content is the default operational practice, with carve-outs for safety-critical and user-initiated reporting scenarios that remain subject to model training independent of user opt-out elections. This structure creates distinct data handling pathways based on content classification and user action.
CA-P-004650 First tracked May 7, 2026 Last seen May 7, 2026 Compare across platforms →
OpenAI · Privacy Policy (ROW)
The clause establishes OpenAI's baseline authority to incorporate user content into model training while providing a mechanism for users to restrict this practice. The provision clarifies that safety review processes operate independently of the model training opt-out, establishing distinct operational pathways for content use.
CA-P-007087 First tracked May 9, 2026 Last seen May 9, 2026 Compare across platforms →
Fireworks AI · Fireworks AI Privacy Policy
This provision establishes the operational scope of data usage for AI model development, clarifying that training purposes require affirmative user authorization rather than occurring as a default practice. The mechanism creates a procedural requirement for consent before a particular category of data processing occurs.
CA-P-005144 First tracked May 7, 2026 Last seen May 7, 2026 Compare across platforms →

Compliance Governance Intelligence

Monitor specific governance provisions across platforms.

Compliance includes provision-level monitoring, regulatory mapping, and audit-ready analysis.

Start free Start Compliance free trial