-
Perplexity AI
· Perplexity Terms of Service
Even if Perplexity causes you harm through its service, the maximum amount the company owes you is either $100 or what you paid in the last 12 months, whichever is higher. You cannot recover lost profits, data loss, or other indirect damages....
Why it matters: The $100 floor on total liability means that in many cases, the maximum financial remedy available to a user against Perplexity is $100, regardless of the actual harm caused, which significantly limits practical recourse for most users....
-
Perplexity AI
· Perplexity Terms of Service
Perplexity's service is not permitted for children under 13, and the company states it will delete personal data if it discovers a child under 13 has used the service....
Why it matters: The prohibition on under-13 use and the stated data deletion commitment engage COPPA compliance obligations, but the terms rely primarily on users self-certifying their age rather than implementing active age verification, which is an area of increasing regulatory scrutiny....
-
Perplexity AI
· Perplexity Terms of Service
Perplexity can suspend or terminate your account at any time without warning, for any reason, at the company's sole discretion. You can also cancel your own account by email or through settings....
Why it matters: The right to terminate without notice or stated cause means that users, including paid subscribers, may lose access to the service and any associated data without prior opportunity to export or migrate their information....
-
Perplexity AI
· Perplexity Terms of Service
If your use of Perplexity causes legal claims against the company, including from third parties claiming your content violated their rights, you agree to pay Perplexity's legal costs and damages....
Why it matters: This clause creates financial exposure for users whose submitted content or use of the service results in third-party legal claims against Perplexity, including claims related to copyright infringement, privacy violations, or defamation....
-
Runway
· Runway Privacy Policy
Everything you input into Runway, including text prompts, uploaded images, videos, audio, and any content generated through the service, is collected along with detailed metadata about when, where, and how that content was created or edited....
Why it matters: The policy authorizes collection of the full scope of user-generated content including creative prompts and AI-generated outputs, along with associated metadata. For professional or enterprise users, this may include proprietary creative work, confidential project details, or sensitive subject matter submitted as prompts....
-
Monitoring
These provisions have changed before.
Monitor includes same-day alerts, structured change summaries, and monitoring for up to 25 platforms.
-
Runway
· Runway Privacy Policy
If you use Runway through a workplace or team account, your employer's account administrator can access and control your Runway account. Runway may also share your email address with your employer if you signed up using a work email address....
Why it matters: Employees using Runway with work email addresses may have their account information disclosed to their employer without a separate consent step, and enterprise account administrators are granted access and control rights over individual user accounts....
-
Runway
· Runway Privacy Policy
If Runway is sold, merged, or goes through a financing event, your personal data may be shared with or transferred to the acquiring company or prospective buyers, including in the event of bankruptcy....
Why it matters: This provision permits disclosure of user personal data, including biometric data and user content, to unaffiliated third parties in connection with corporate transactions, including to prospective counterparties before any transaction is finalized. The scope of data transferable as a business asset is not specifically limited....
-
Runway
· Runway Privacy Policy
For EU and UK users, Runway relies on 'legitimate interests' as its legal basis for a broad set of data processing activities including analytics, business improvement, security, and fraud prevention. This means Runway has assessed that its business interests outweigh user privacy interests for these activities....
Why it matters: Legitimate interests is one of the broadest GDPR legal bases and is used here to cover analytics, service improvement, security, and fraud prevention. EU and UK users have the right to object to processing carried out under legitimate interests, which could limit how Runway processes their data....
-
Runway
· Runway Privacy Policy
Runway's service is not intended for children under 13, and Runway states it does not knowingly collect personal data from this age group. Parents who believe a child under 13 has submitted data can contact Runway to request deletion....
Why it matters: This provision establishes Runway's COPPA compliance posture, relying on a 'not directed to children' standard. The policy does not describe any age verification mechanism, which is a common limitation in online service COPPA compliance. The collection of biometric data from users raises heightened considerations if any underage users access the service....
-
Runway
· Runway Privacy Policy
Your personal data may be stored and processed in the United States or other countries that may have different or less protective privacy laws than your home country. Using the service is treated as acknowledgment of this transfer....
Why it matters: For EU and UK users, international data transfers to the United States require a lawful transfer mechanism under GDPR Chapter V, such as Standard Contractual Clauses. The policy's reliance on user acknowledgment through service use as a consent mechanism for transfers may not satisfy GDPR transfer requirements....
-
Runway
· Runway Terms of Service
Your Runway subscription renews automatically at the end of each billing period at whatever price Runway is currently charging, and your payment method is charged without additional notice unless you cancel first....
Why it matters: The agreement authorizes automatic charges at Runway's then-current pricing, meaning price increases applied before a renewal date will be charged without requiring separate user affirmation beyond continued use, and no prorated refund is provided if you cancel mid-period....
-
Runway
· Runway Terms of Service
When you post content publicly on Runway, you grant the company a broad, permanent license to reproduce, modify, distribute, and make money from that content in any current or future media format....
Why it matters: The license granted over publicly posted content is stated as irrevocable and perpetual, covering derivative works and revenue-generating uses, which means Runway retains these rights even if you delete the content or close your account....
-
Runway
· Runway Terms of Service
You are not permitted to use Runway's services or the content they generate to build, train, or improve AI tools or other products that compete with Runway....
Why it matters: This restriction applies to AI-generated Outputs and prohibits their use in developing competitive products, which may limit how developers, researchers, and businesses can use content generated through Runway's platform....
-
Runway
· Runway Terms of Service
Runway can change these terms at any time, and your continued use of the service after changes are posted counts as your agreement to the new terms....
Why it matters: The agreement permits Runway to modify terms unilaterally at any time with website posting as the sole required notice, and continued use constitutes acceptance, which means users may be bound by updated terms without receiving direct notification....
-
Runway
· Runway Terms of Service
By using Runway's services, you agree that your uploaded content, chat messages, and voice communications are not private, and that Runway may review this content for its own purposes....
Why it matters: The agreement states users have no expectation of privacy regarding transmitted content, including chat and voice communications, which is relevant to how users assess what information they share through the platform....
-
Pika
· Pika Acceptable Use Policy
Pika states that everything you upload to the service and every video it generates for you is your legal responsibility, including any risks that arise from sharing that content....
Why it matters: This provision asserts that users bear full legal responsibility for the accuracy, quality, legality, and appropriateness of all inputs and outputs, which may have significant implications if AI-generated content infringes third-party rights or causes harm....
-
Pika
· Pika Acceptable Use Policy
If you create a deepfake or AI-manipulated video using Pika, you are required to disclose to viewers that the content was artificially generated or manipulated....
Why it matters: This provision assigns users the obligation to disclose AI-generated or manipulated outputs, which aligns with and incorporates by reference applicable state and federal deepfake disclosure laws, placing legal compliance responsibility on users rather than the platform....
-
Pika
· Pika Acceptable Use Policy
Pika can remove your content or suspend your access at any time if it decides, entirely on its own, that your content poses a risk, even if nothing you did is specifically listed as prohibited....
Why it matters: This provision grants Pika authority to take enforcement action against content or users based on its own discretionary assessment, without requiring that the conduct fall within any enumerated prohibition, which means account actions are not limited to the specific rules listed in the AUP....
-
Pika
· Pika Acceptable Use Policy
Pika can share your account information and identity with law enforcement at any time, and by using the service you agree not to sue Pika if you are harmed as a result of that disclosure or any investigation....
Why it matters: This provision asserts Pika's authority to disclose user identity and information to law enforcement in response to requests or court orders, and requires users to indemnify Pika against claims arising from such disclosures or related investigations; the enforceability of the indemnification waiver may vary by jurisdiction....
-
Pika
· Pika Acceptable Use Policy
You are not allowed to use videos or other content created by Pika's AI to build a competing AI video service or product....
Why it matters: This provision restricts downstream commercial use of AI-generated outputs, prohibiting users from using outputs to develop or offer competing products or services, which may have significant implications for developers, researchers, and enterprise users....
-
Pika
· Pika Acceptable Use Policy
You cannot use Pika to create political ads, campaign materials, or any content intended to interfere with elections or political processes....
Why it matters: This provision explicitly prohibits using the service for political advertising or election interference, which addresses a recognized risk area for AI-generated content in the context of democratic processes....
-
Suno
· Suno Acceptable Use Policy
Suno's page code sets advertising and analytics tracking to 'off' by default for users in EU, EEA, UK, Switzerland, Norway, Iceland, and Liechtenstein, but sets all tracking to 'on' by default for everyone else, including US users....
Why it matters: This configuration determines whether advertising identifiers and behavioral analytics are collected from your device before you interact with any consent prompt, and the default-granted posture for non-EU users means tracking begins without an explicit opt-in....
-
Suno
· Suno Acceptable Use Policy
The ability to use Suno-generated music commercially is only available on the Pro plan ($10/month) and above; the free tier does not include commercial rights....
Why it matters: Free-tier users who use Suno-generated music in commercial projects, monetized videos, advertising, or other revenue-generating contexts may be doing so outside the scope of their licensed rights under the terms as disclosed in the structured data....
-
Suno
· Suno Acceptable Use Policy
Suno's homepage loads tracking scripts from Microsoft Clarity, TikTok, Meta (Facebook), Bing Ads, and Twitter, which may collect device identifiers, browsing behavior, and interaction data from visitors....
Why it matters: Each of these third-party scripts transmits data about your visit to separate companies (Microsoft, ByteDance/TikTok, Meta, Twitter), not just to Suno, and this occurs under the default-granted consent configuration for non-EU users....
-
Suno
· Suno Acceptable Use Policy
Suno offers annual subscription plans at $96/year (Pro) and $288/year (Premier); these are billed as lump-sum annual charges, and the full terms governing auto-renewal, cancellation, and refund rights are not visible in the submitted HTML....
Why it matters: Annual subscription billing involves a single upfront charge of up to $288, and users need to understand the cancellation and refund policy before committing, as these terms are not disclosed in the pricing structured data alone....
-
Hugging Face
· Hugging Face Model Card Guidelines
Hugging Face requires model publishers to include structured metadata in their model cards, covering license type, language, and task type, and recommends additional fields for training data, evaluation results, and limitations....
Why it matters: The metadata fields establish the baseline disclosure a model publisher must provide, which directly affects how users understand what they are permitted to do with a model and what risks it may carry....
-
Hugging Face
· Hugging Face Model Card Guidelines
Model card authors are encouraged to document known biases and limitations of their AI models, so that users can make informed decisions about whether and how to use them....
Why it matters: Bias and limitations disclosures are directly relevant to responsible AI deployment decisions, particularly in regulated contexts such as hiring, lending, healthcare, or law enforcement, where algorithmic bias may create legal liability....
-
Hugging Face
· Hugging Face Model Card Guidelines
Model publishers are encouraged to state explicitly what their model is designed to do and what uses they consider inappropriate or outside the model's intended purpose....
Why it matters: Intended use declarations define the scope of responsible use as understood by the model publisher, and out-of-scope use declarations may be relevant to liability allocation if a model is deployed in a context the publisher explicitly identified as inappropriate....
-
Hugging Face
· Hugging Face Model Card Guidelines
Model publishers may include structured performance evaluation data in model cards, specifying which benchmarks were used, what datasets were evaluated, and what scores the model achieved....
Why it matters: Evaluation results disclosures allow users to assess model performance claims against specific benchmarks, which is material for organizations that need to validate AI model performance before deployment in regulated or high-stakes contexts....
-
Hugging Face
· Hugging Face Model Card Guidelines
Model publishers are encouraged to disclose what datasets were used to train their model, which helps users assess potential biases, data provenance issues, and licensing implications of the training data....
Why it matters: Training data disclosure is directly relevant to intellectual property compliance, data provenance assessments, and bias risk evaluation, particularly as regulatory frameworks increasingly require transparency about AI training data sources....
-
DeepSeek
· DeepSeek Model License
Businesses with more than 100 million monthly active users must get written permission from DeepSeek before using the model commercially; those below this threshold may use it commercially without separate approval....
Why it matters: This provision creates a tiered commercial licensing structure where large-scale platforms face an additional contractual barrier to deployment, requiring affirmative consent from DeepSeek before commercial use, which introduces dependency on DeepSeek's licensing decisions for those entities....
-
DeepSeek
· DeepSeek Model License
DeepSeek grants you a free, non-transferable, revocable license to use, modify, and distribute the model for research or commercial purposes, subject to the conditions set out in the license....
Why it matters: The license is explicitly revocable, meaning DeepSeek retains the right to terminate the license grant, which creates ongoing dependency on DeepSeek's continued authorization for any organization that has built commercial products on the model....
-
DeepSeek
· DeepSeek Model License
DeepSeek provides the model with no guarantees about its performance, accuracy, or fitness for any specific use; if the model causes harm or fails to work as expected, DeepSeek is not liable under the license....
Why it matters: This clause means that organizations deploying DeepSeek-R1 in commercial products cannot rely on any warranty from DeepSeek regarding model quality, safety, or fitness for purpose, and must independently validate the model's performance for their specific use case....
-
DeepSeek
· DeepSeek Model License
DeepSeek cannot be held liable for indirect or consequential damages arising from use of the model, even if such damages result from reliance on the model's outputs....
Why it matters: This clause limits the financial exposure DeepSeek faces if the model causes harm, meaning that organizations or individuals harmed by model outputs would need to seek recourse from the deploying organization rather than from DeepSeek directly....
-
ElevenLabs
· ElevenLabs Privacy Policy
ElevenLabs shares user data including identifiers and behavioral data with advertising and analytics companies to measure service usage and serve targeted advertising....
Why it matters: Sharing personal data with advertising and analytics partners may constitute a sale or sharing of personal information under CCPA, triggering opt-out rights for California residents, and requires valid consent under GDPR for cookie-based tracking....
-
ElevenLabs
· ElevenLabs Privacy Policy
Users in the EU, UK, and Switzerland can ask ElevenLabs to access, correct, delete, or export their personal data by contacting privacy@elevenlabs.io....
Why it matters: These rights are enforceable under GDPR and UK GDPR, and ElevenLabs is obligated to respond to valid requests within statutory timeframes; failure to honor these rights can be reported to national data protection authorities....
-
ElevenLabs
· ElevenLabs Privacy Policy
ElevenLabs states its services are not intended for children under 13, and the company states it will delete any personal data collected from a child under 13 if discovered....
Why it matters: This provision engages COPPA obligations for US users; the policy does not describe age verification mechanisms, which is relevant given that voice cloning tools could be accessed by minors....
-
ElevenLabs
· ElevenLabs Privacy Policy
When you use ElevenLabs from outside the US, your personal data including voice recordings is transferred to and stored in the United States, where data protection laws may differ from your home country....
Why it matters: Cross-border transfers of personal data from the EEA or UK to the US require a valid transfer mechanism under GDPR Chapter V, such as Standard Contractual Clauses; the policy acknowledges transfer but does not specify the legal mechanism used....
-
ElevenLabs
· ElevenLabs Privacy Policy
ElevenLabs keeps your personal data for as long as it considers necessary for the stated purposes, without specifying fixed retention periods for particular data types such as voice recordings....
Why it matters: The absence of specific retention periods for voice recordings and other sensitive data categories means users cannot know how long this data will be held; GDPR's storage limitation principle and some state biometric statutes require defined retention limits....
-
ElevenLabs
· ElevenLabs Privacy Policy
If ElevenLabs is sold, merges with another company, or undergoes a similar corporate event, your personal data including voice recordings may be transferred to the acquiring entity....
Why it matters: This provision means that user data, including voice recordings and audio content, may be transferred to a third party in a corporate acquisition without requiring individual user consent, subject to applicable law....
-
ElevenLabs
· ElevenLabs Privacy Policy
ElevenLabs and its partners use cookies and tracking tools to collect your browsing activity, device identifiers, and IP address for analytics and advertising purposes....
Why it matters: Cookie-based tracking for advertising purposes requires informed consent under GDPR and the ePrivacy Directive for EEA and UK users; California residents have CCPA rights regarding personal data collected through tracking technologies....
-
ElevenLabs
· ElevenLabs Safety Policy
You cannot use ElevenLabs to create fake audio recordings that spread false information, such as fabricating a statement by a real person that they never made....
Why it matters: The policy specifically names disinformation as a prohibited use, which is relevant to elections, journalism, and public discourse contexts where synthetic audio could be used to fabricate statements by real individuals....
-
ElevenLabs
· ElevenLabs Safety Policy
ElevenLabs has a team that monitors AI-generated content using both software and human reviewers. The company can remove content, shut down your account, and report you to law enforcement if it finds a violation....
Why it matters: The policy authorizes ElevenLabs to suspend or terminate accounts and refer users to law enforcement without specifying the procedural rights available to users in connection with these actions, which is relevant to both individual users and enterprise customers who depend on platform access....
-
ElevenLabs
· ElevenLabs Safety Policy
You cannot use ElevenLabs to create fake voice recordings to commit fraud, steal someone's identity, or trick people or systems for financial benefit....
Why it matters: AI voice synthesis has been used in documented fraud cases including CEO impersonation scams and telephone authentication bypasses; this provision addresses that risk directly and signals ElevenLabs' position on law enforcement cooperation....
-
Cursor
· Cursor Security Practices
When Privacy Mode is on, Cursor contractually requires its AI model providers not to store your code or use it to train their models. This mode is on automatically for team members and can be turned on manually by any individual user....
Why it matters: The document states that without Privacy Mode, code data may be stored or used for training by model providers; enabling it triggers contractual ZDR obligations on those third-party providers....
-
Cursor
· Cursor Security Practices
Cursor accepts vulnerability reports by email and acknowledges them within 5 business days; critical security incidents are communicated by email to affected users....
Why it matters: The document states that critical incidents will be communicated to affected users by email, which is the mechanism users should expect to receive breach or incident notifications; users should ensure their email address on file is current....
-
Stability AI
· Stability AI Model License
Organizations whose annual revenue exceeds a specified threshold must obtain a paid commercial license from Stability AI to use its models, rather than relying on the free community license available to smaller or non-commercial users....
Why it matters: This provision determines whether a business must pay for a license, and the threshold creates a compliance trigger that organizations need to monitor as they grow....
-
Stability AI
· Stability AI Model License
The license grants qualifying users the right to download and run Stability AI model weights on their own infrastructure, rather than requiring use through Stability AI's API, subject to the applicable license tier and acceptable use obligations....
Why it matters: Self-hosting rights distinguish this license from API-only AI service agreements and create a distinct set of obligations for deployers who control the model infrastructure directly....
-
Stability AI
· Stability AI Model License
Stability AI retains ownership of the model weights and associated intellectual property; the license grants only a limited right to use the models under specified conditions, not a transfer of ownership....
Why it matters: This provision establishes that licensees do not own the model weights they download and self-host, meaning Stability AI can modify or revoke license terms for future versions and enforce IP rights against unauthorized uses....
-
Stability AI
· Stability AI Model License
Organizations that deploy Stability AI models in their own products are required to pass through acceptable use obligations to their own users, meaning end users of third-party applications built on these models are also bound by Stability AI's use restrictions....
Why it matters: This provision creates a compliance obligation for deployers to implement and enforce acceptable use terms with their own customers, extending Stability AI's policy framework through the distribution chain....