Vercel

Summary

The diff only states that approximately 44 words were removed; it does not identify the deleted language.

AI Training and Data Use

• No determination is possible regarding whether the removed text changed:
• The provider’s right to use customer data to train, fine-tune, or improve AI models;
• Whether customer data may be used for training by default or only with consent;
• Whether inputs, outputs, prompts, or usage data are included in “customer data”;
• Restrictions on using confidential, personal, or regulated information;
• Opt-out, deletion, retention, or anonymization rights; or
• Whether subcontractors or affiliates may use the data for AI training.

Potential Legal Significance

• Removing language could either reduce customer protections—for example, by deleting a prohibition or opt-out right—or reduce the provider’s rights, such as by deleting a broad data-use authorization.
• The legal effect cannot be assessed without the actual 44 deleted words and the surrounding provisions. Context is particularly important because deletion may alter definitions, exceptions, consent requirements, or the scope of a data-use license.

Recommended Follow-Up

Obtain the full redline or the text immediately before and after the deletion. Review specifically for changes to:

1. Customer-data ownership and permitted uses;

2. AI model training, improvement, or product development;

3. Confidentiality and personal-data restrictions;

4. Data retention, deletion, and de-identification;

5. Opt-out or consent mechanisms; and

6. Liability and indemnity for unauthorized data use.

Conclusion: The provided diff is insufficient to identify any specific contractual change or new risk, including changes concerning AI-model training.

Key Changes and Risks

1. Customer content may be used for AI model training

• The revised terms expressly state that Vercel may use Your Content to train Vercel’s artificial intelligence and machine-learning models.
• Vercel may also share Your Content with third-party products or providers, including for training and improving their AI and machine-learning models.
• The license supporting this use is broad: worldwide, royalty-free, fully paid, sublicensable, transferable, and permits copying, modification, adaptation, reproduction, distribution, publication, storage, performance, and creation of derivative works.
• The stated purposes include:
• Providing and improving the Services;
• Developing new products and services;
• Securing and protecting the Services and third parties; and
• Preventing fraud, abuse, malware, and malicious content.
• The text indicates that Model Training is not permitted for customers on a paid Pro plan, but the precise scope of that exclusion is not fully visible in the diff. The terms should be checked to determine whether the exclusion is automatic, applies only to certain content, or depends on an election or setting.
• The revised terms do not appear to provide a general no-training right for free, Hobby, or other plans. Customers using those plans should assume that submitted content may be used for model training unless another incorporated policy says otherwise.

Risk

Customers may unintentionally submit confidential information, source code, personal data, regulated data, trade secrets, or third-party materials that can be used to train models or shared with AI providers. The customer represents that it has sufficient rights for Vercel’s broad use of the content, potentially increasing infringement, confidentiality, privacy, and regulatory exposure.

2. Broader and clearer rights in customer content

• The prior technical-support access language has been replaced or expanded into a broad license to use Your Content for service improvement, product development, security, and AI-related purposes.
• The license is transferable and sublicensable, allowing use by affiliates, contractors, and third-party service providers.
• The customer remains responsible for all content and activities under its account.

3. Third-party AI tools and autonomous actions

• The terms now expressly address AI functionality, AI providers, agents, and third-party tools.
• If enabled, third-party AI tools may access the Services and take actions on the customer’s behalf.
• Customers are responsible for costs and actions attributable to those tools, which are treated as the customer’s electronic agents.

4. Dispute-resolution changes and emphasis

• The agreement adds prominent notice of mandatory binding arbitration and a class-action waiver.
• Claims generally must be brought individually, subject to stated exceptions.
• A 30-day opt-out process is provided.
• California law and San Francisco-related arbitration procedures apply.

5. Other notable changes

• The agreement is reorganized and incorporates additional terms, including AI Product Terms, DORA, API, Marketplace, and other service-specific addenda.
• Account-security obligations are more explicit: unique credentials, no sharing, accurate signup information, and responsibility for account activity.
• Vercel retains broad ownership of System Data and may change service limits and practices at its discretion.

Summary

The diff states only that approximately 77 words were removed from the document. The actual deleted language is not provided.

Legal and Commercial Impact

Because the removed text is unavailable, it is not possible to determine:

• Which contractual rights or obligations were deleted;
• Whether liability, indemnity, confidentiality, security, or termination provisions changed;
• Whether customer data rights were narrowed or expanded;
• Whether any consent, notice, audit, or compliance obligations were removed; or
• Whether the deletion creates ambiguity or shifts risk to the customer.

AI Training and Data Use

No specific conclusion can be reached regarding the use of customer data to train AI models. The removed language could potentially have addressed:

• Whether customer data, prompts, inputs, outputs, or usage data may be used to train or improve AI models;
• Whether data may be used for general model training versus service-specific fine-tuning;
• Whether customer consent is required;
• Whether data is anonymized, aggregated, or de-identified before use;
• Whether human reviewers or third-party AI providers may access the data;
• Whether customers may opt out of training or revoke permission; or
• Whether training rights continue after termination.

If the deleted language contained a customer-data license or AI-training permission, its removal could either reduce the provider’s rights or eliminate important limitations and safeguards. Without the actual text, the direction and significance of the change cannot be determined.

Recommended Follow-Up

Obtain the exact 77 deleted words, preferably in a redline showing the surrounding provisions. Review particularly any language containing terms such as “train,” “improve,” “develop,” “machine learning,” “artificial intelligence,” “inputs,” “outputs,” “usage data,” “de-identified,” “aggregated,” “license,” or “service improvement.”

Key Changes and Risks

1. Customer Content may be used for AI training

The most significant change is the addition of a Model Training provision. Customers agree that Vercel may:

• Use Your Content to train Vercel’s artificial-intelligence and machine-learning models; and
• Share Your Content with third parties, including for training and improving their AI and machine-learning models.

The provision appears to define this broadly and does not, in the supplied diff, limit the use to anonymized, aggregated, or de-identified data. “Your Content” includes software, code, data, text, inputs, feedback, and other materials submitted to or stored through the Services.

Risks

• Confidential source code, proprietary information, personal data, or regulated data could be used in model development or disclosed to AI providers.
• The third-party sharing right may allow onward processing outside Vercel’s direct control.
• The language does not clearly state retention periods, deletion procedures, geographic restrictions, or whether trained models can be deleted or unlearned.
• Customers may need to obtain consent from employees, clients, or data subjects before submitting information.
• The interaction between this permission, the DPA, privacy laws, confidentiality obligations, and sector-specific requirements is unclear.

The diff also states that Model Training is not available if the customer is on a paid Pro plan. The exact scope and operation of that apparent opt-out or plan-based exclusion should be confirmed. It is unclear whether it applies automatically, covers all Content, or extends to third-party AI providers.

2. Broader Content license

The customer license to Vercel is expressly stated to be worldwide, royalty-free, fully paid, sublicensable, and transferable. It covers copying, modifying, reproducing, distributing, displaying, publishing, storing, performing, and creating derivatives of Content.

The stated purposes now include providing and improving Services, developing new products and services, security, and protecting against fraud, abuse, malware, and malicious files. Although these purposes are not expressly limited to AI training, they reinforce Vercel’s broad operational and product-development rights.

3. New or expanded dispute restrictions

The revised terms prominently require binding arbitration and waive class, collective, multiple-plaintiff, and representative actions, subject to stated exceptions. Customers generally have only 30 days to opt out. Disputes are governed by California law and are generally handled through JAMS, with San Francisco identified as the venue.

4. Additional contractual and operational changes

• The terms incorporate numerous external policies and addenda, including AI Product Terms, the DPA, Acceptable Use Policy, API terms, Marketplace terms, and DORA terms.
• Customers remain responsible for all activity under their accounts and must use unique, non-shared credentials.
• Vercel may access accounts to provide technical support.
• Vercel retains ownership of System Data, including telemetry, logs, traffic data, and usage statistics.
• Vercel receives broader rights to customer feedback and permission to use customer names, trademarks, and logos in marketing materials.

Summary of Available Changes

The diff states only that approximately 77 words were removed. The actual deleted language is not provided.

AI Training and Data Use

• No determination is possible regarding whether the removed text addressed:
• Use of customer data to train, fine-tune, or improve AI models;
• Whether customer data may be used for generalized model training;
• De-identification, aggregation, or anonymization requirements;
• Customer consent or opt-out rights;
• Restrictions on using prompts, inputs, outputs, or uploaded content;
• Ownership or licensing of data used for AI purposes; or
• Retention and deletion of data used in AI systems.

• If the deleted language contained restrictions or customer protections relating to AI training, its removal could materially expand the provider’s rights to use customer data or create ambiguity about permitted uses.

Other Potential Legal Risks

Because the deleted words are not shown, it is not possible to assess whether the amendment changes:

• Confidentiality obligations;
• Data-processing or privacy commitments;
• Security requirements;
• Intellectual-property ownership or licenses;
• Liability, indemnity, or warranty exposure;
• Data retention and deletion obligations;
• Subprocessor or third-party use rights; or
• Termination and post-termination data handling.

Recommended Action

Obtain the precise redlined text or a version showing the deleted language. The impact of a deletion depends entirely on what was removed; a 77-word deletion could be immaterial or could eliminate a significant customer protection, particularly if it concerned AI training or secondary use of customer data.

Key Changes and Risks

1. Customer Content and AI Model Training

• The revised terms retain a broad license to Customer Content. Vercel may use, copy, modify, reproduce, distribute, display, publish, store, perform, and create derivative works from the content.
• The stated purposes are expanded to include:
• Providing and improving the Services;
• Developing new products and services;
• Securing and protecting the Services; and
• Protecting Vercel and third parties from fraud, abuse, malware, and malicious content.
• The terms expressly state that Vercel may use Customer Content to train its artificial intelligence and machine-learning models and may share Customer Content with third-party products, including for training and improving their AI and machine-learning models.
• A potentially important limitation is added: if the customer is on a paid Pro plan, “Model Training” is not [permitted]. The excerpt does not show the rest of that sentence or explain:
• Whether the restriction applies automatically or requires an opt-out;
• Whether it applies to all Customer Content or only content submitted through particular features;
• Whether it covers Vercel’s affiliates and third-party providers;
• Whether it applies to previously collected content; or
• Whether it applies to account, system, telemetry, or de-identified data.
• The terms therefore create a significant risk for customers on free, Hobby, enterprise, or other non-Pro plans: their content may be used for AI training and shared with third parties unless another provision limits that use.
• Customers should confirm the complete Model Training provision and any applicable AI Product Terms, DPA, privacy policy, and plan-specific terms.

2. Broader Content License

• The license remains worldwide, royalty-free, fully paid, sublicensable, transferable, and non-exclusive.
• It applies to broadly defined content, including code, software, data, text, feedback, inputs, and other materials.
• Customers continue to bear responsibility for ensuring they have the necessary rights and consents. This is particularly important where content contains personal information, confidential business information, regulated data, or third-party intellectual property.

3. AI Features and Third-Party Tools

• New or more clearly organized provisions address AI Functionality and third-party AI tools.
• AI features may take autonomous actions. Customers are responsible for those actions and for costs incurred by third-party tools using the Services.
• Third-party tools are treated as the customer’s “electronic agent,” potentially making the customer responsible for their acts, instructions, and resulting use of the platform.
• AI Products and Services are also made subject to separate AI Product Terms incorporated by reference.

4. Dispute Resolution

• The arbitration and class-action waiver are prominently emphasized and incorporated into the acceptance language.
• Claims generally must be resolved through individual, binding arbitration, with class and representative proceedings waived to the fullest extent permitted by law.
• A 30-day opt-out procedure is specified. Failure to opt out promptly may materially limit litigation options.

5. Other Notable Changes

• Open-source software, including Next.js, is expressly excluded from the Agreement and governed by its own licenses.
• The terms more clearly state that Services may be supplied by Vercel affiliates.
• Account security obligations are reinforced, including unique credentials for each user and responsibility for account activity.
• Additional incorporated terms now expressly include AI Products, DORA, API, CDN, Marketplace, domain registration, and other service-specific addenda.

Summary

The provided diff states only that approximately 77 words were removed from the document. It does not identify:

• Which provisions were deleted;
• Whether the deletions were replaced with different language;
• Whether any data-use, confidentiality, privacy, intellectual-property, or AI-related terms changed; or
• Whether the customer’s data may be used to train, improve, fine-tune, evaluate, or develop AI models.

Legal and Commercial Impact

The impact cannot be reliably assessed from the information provided. Deleting language could:

• Remove restrictions on the provider’s use of customer data;
• Eliminate consent, notice, or opt-out requirements;
• Narrow confidentiality or security obligations;
• Change ownership or licensing rights in customer data or outputs;
• Remove limitations on using data for analytics, product improvement, or machine-learning purposes; or
• Remove protections requiring deletion or return of data.

However, none of these changes can be confirmed without seeing the actual deleted text and any replacement language.

AI Training/Data-Use Review

No specific change concerning AI model training can be identified from the abbreviated diff. The available description does not establish whether customer data may be used to:

• Train or fine-tune generative AI or other models;
• Improve or develop the provider’s products;
• Create aggregated, de-identified, or derived datasets;
• Share data with affiliates, vendors, or model providers; or
• Retain data for model-training purposes after termination.

Required Information

Please provide the actual redline, including the 77 deleted words and any surrounding provisions. The most important clauses to review are those addressing:

1. Customer data and content;

2. Confidentiality and privacy;

3. Product improvement or analytics;

4. Artificial intelligence or machine learning;

5. Data retention and deletion;

6. Ownership and licenses; and

7. Subprocessors or third-party AI providers.

Summary

The provided diff states only that approximately 77 words were removed. It does not identify which provisions were deleted or provide the replacement text.

AI Training and Data Use

• No determination is possible regarding whether the changes affect:
• The customer’s data being used to train AI models;
• Whether training is permitted by default or requires consent;
• Use of customer content to improve products or services;
• Anonymization, aggregation, or de-identification requirements;
• Opt-out or deletion rights;
• Use of personal data, confidential information, or uploaded content for model development; or
• Restrictions on human review or third-party access.

Because the deleted language is not provided, removal of the 77 words could either expand or restrict the provider’s rights. In particular, deleting an AI-training limitation, consent requirement, or opt-out mechanism could materially increase the customer’s privacy and confidentiality risk.

Other Legal Risks

The available information is insufficient to assess whether the deletions affect:

• Ownership or licensing of customer data;
• Confidentiality obligations;
• Data retention and deletion;
• Security commitments;
• Liability, indemnities, or warranties;
• Compliance with privacy laws;
• Subprocessors or data transfers; or
• Termination rights.

Conclusion

No reliable legal comparison can be made from the description alone. The actual deleted text and any surrounding provisions are needed to determine the effect of the amendment, particularly on AI model training and use of customer data.

Summary of Important Changes and Risks

1. Customer Content may be used for AI training

The revised terms expressly state that Vercel may use “Your Content” to train and improve artificial intelligence and machine-learning models and may share Your Content with third-party products, including for training and improving their AI models.

Key risks

• This is a significant expansion or clarification of Vercel’s rights to use customer-provided content.
• “Your Content” is broadly defined to include software, source code, data, information, feedback, text, inputs, and other materials submitted to or stored through the Services.
• The license remains worldwide, royalty-free, transferable, sublicensable, and permits copying, modification, reproduction, distribution, publication, performance, and creation of derivative works.
• The stated purposes now expressly include providing and improving the Services, developing new products and services, security, fraud prevention, abuse prevention, malware prevention, and AI/model training.
• Customers should confirm that they have the necessary rights, consents, and confidentiality permissions to submit data for these purposes, particularly personal data, regulated information, proprietary code, and third-party content.
• The diff references an exception for customers on a paid Pro plan (“Model Training is not…”), but the excerpt does not show the complete sentence. The scope, default status, and opt-out mechanics should be verified in the final version. No equivalent protection is apparent for all other plans.

2. Expanded use and access rights

Vercel retains an express right to access a customer account to provide technical support. The customer grants Vercel broad rights to use submitted content, including to develop new products and services. This may permit uses beyond merely hosting or operating the customer’s deployment.

3. AI functionality and third-party tools

The revised terms more clearly address AI-enabled features, autonomous actions, and third-party AI tools. Customers are responsible for actions taken by those tools on their behalf and for costs incurred through their use. This creates operational and authorization risks where AI agents can modify projects, access data, or incur usage charges.

4. Dispute resolution

The arbitration and class-action waiver provisions are prominently emphasized and reorganized. The agreement requires binding arbitration, generally waives class, representative, and multi-plaintiff claims, and provides a 30-day opt-out process. Claims involving intellectual property and certain injunctive relief are exceptions.

5. Other material changes

• The agreement more clearly applies when accepted by a company, with the accepting person representing authority to bind it.
• Services now expressly include products offered by Vercel affiliates.
• Open-source software, including Next.js, is excluded from the agreement and governed by its applicable licenses.
• Data protection language continues to reference the DPA and Privacy Policy, while Vercel retains ownership of System Data.
• Vercel receives a license to use the customer’s name, trademarks, and logos in marketing materials, subject to the stated terms.

Structured Summary of Important Changes

1. Expanded AI and Model-Training Rights — High Risk

The revised terms expressly state that Vercel may use Your Content to train artificial intelligence and machine-learning models. They also permit Vercel to share Your Content with third-party products, including for training and improving those parties’ AI and machine-learning models.

This is a significant expansion of data-use rights because:

• “Your Content” includes software, source code, data, text, inputs, feedback, and other materials submitted to the Services.
• The permission is structured as a broad, worldwide, royalty-free, sublicensable and transferable license.
• The stated purposes include providing and improving the Services, developing new products and services, security and fraud prevention, and third-party AI/model training.
• The excerpt indicates a potential limitation for customers on a paid Pro plan (“Model Training is not…”), but the provided diff does not include the complete sentence. The scope, default setting, opt-out process, and applicability of this exception should be confirmed in the complete agreement.
• There is no apparent restriction in the excerpt limiting training to aggregated, anonymized, or de-identified data.

Customer risk: Confidential source code, personal information, proprietary data, or regulated information could potentially be used to develop models or shared with third-party AI providers unless an applicable plan-based exclusion, opt-out, DPA provision, or separate AI Product Term limits that use.

2. Broader License to Customer Content

The license to Your Content is clarified or expanded to cover copying, modifying, adapting, reproducing, distributing, displaying, publishing, storing, performing, and creating derivatives. Vercel may use the content to develop products, secure the Services, and protect Vercel and third parties from fraud, abuse, malware, and malicious files.

Customers should verify that they have authority to grant these rights, especially for third-party data, employee information, customer data, open-source materials, and confidential code.

3. AI Functionality and Third-Party Tools

A new or more prominent section addresses AI features, AI providers, autonomous actions, and third-party tools. Customers are responsible for:

• Actions taken by AI functionality or third-party tools on their behalf;
• Costs incurred through third-party tools’ use of the Services; and
• The legal effect of treating those tools as an electronic agent under UETA or similar law.

This creates operational, authorization, and liability risks where AI agents can deploy code, modify projects, or take other autonomous actions.

4. Dispute Resolution

The terms add prominent binding arbitration and class-action waiver language, governed by California law and generally requiring individual JAMS arbitration. A 30-day opt-out is provided, but the customer must send timely written notice.

5. Other Notable Changes

• More detailed account-security obligations, including unique credentials for each user.
• Additional incorporated terms, including AI Product Terms, DPA, Acceptable Use Policy, and other service-specific addenda.
• Vercel may access accounts for technical support.
• Customers remain responsible for account activities and compliance with applicable policies and laws.

Summary

Key Change

• The diff states that approximately 77 words were removed from the document.
• The actual deleted text is not provided, so the nature and legal effect of the changes cannot be determined.

Customer Data and AI Training

• No conclusions can be drawn about whether the agreement now permits, restricts, or changes the use of customer data to:
• Train, fine-tune, or improve AI models;
• Develop products or services;
• Create aggregated, anonymized, or de-identified datasets; or
• Share data with affiliates, vendors, or model providers.
• The deleted language could have contained important limitations, consent requirements, ownership protections, confidentiality obligations, or opt-out rights related to AI training.

Potential Risks

Because the removed wording is unavailable, the following risks cannot be assessed:

• Whether customer data may now be used for AI training without additional consent;
• Whether deletion of restrictions broadens the provider’s rights;
• Whether deletion of a customer-protective clause weakens confidentiality, data-use, or security protections;
• Whether any customer opt-out, deletion, or objection rights were removed;
• Whether the agreement’s treatment of personal data, confidential information, or derived data has changed; and
• Whether liability, indemnity, or regulatory compliance protections were affected.

Required Information

To perform a meaningful legal comparison, provide the actual text of:

1. The deleted approximately 77 words; and

2. The surrounding provision, preferably showing the prior and revised versions.

Without the underlying text, the only reliable conclusion is that material language was removed, but its legal significance—including any change to AI-model training rights—cannot be determined.

Summary

The diff only states that approximately 77 words were removed from the document. It does not identify which provisions were deleted or provide the surrounding language.

AI Training and Data Use

• No conclusions can be drawn about whether the customer’s data may now be used to train, fine-tune, evaluate, or improve AI models.
• The deletion could potentially remove:
• A prohibition on using customer data for AI training;
• A requirement to obtain customer consent before such use;
• Restrictions on using personal, confidential, or regulated data;
• Promises to de-identify, aggregate, or delete data;
• Limits on sharing data with model providers or other subprocessors; or
• Customer audit, opt-out, or deletion rights.
• Conversely, the deleted language could have authorized AI training, in which case its removal might narrow the provider’s rights. The direction of the change cannot be determined from the information provided.

Other Legal Risks

Because the deleted words are not supplied, it is not possible to assess changes to:

• Ownership or licensing rights;
• Confidentiality obligations;
• Privacy and data-protection compliance;
• Security commitments;
• Data retention and deletion;
• Liability, indemnification, or warranties;
• Subprocessor rights;
• Service restrictions; or
• Termination and post-termination obligations.

Recommended Review

The actual deleted text and the corresponding revised clause should be reviewed before concluding that the change is material. Particular attention should be given to any deleted references to “improve,” “develop,” “train,” “fine-tune,” “machine learning,” “artificial intelligence,” “aggregated data,” or “de-identified data.”

Key Changes and Risks

1. Customer Content may be used for AI model training

The revised terms expressly state that Vercel may use Your Content to train artificial intelligence and machine-learning models. “Your Content” includes software, code, data, text, feedback, inputs, and other materials submitted to or stored through the Services.

The license granted to Vercel is broad: worldwide, royalty-free, fully paid, transferable, sublicensable, and includes rights to use, copy, modify, reproduce, distribute, display, publish, store, perform, and create derivative works.

The stated purposes now include:

• Providing and improving the Services;
• Developing new products and services;
• Securing and protecting the Services;
• Protecting Vercel and third parties from fraud, abuse, malware, and malicious content; and
• Training and improving AI and machine-learning models.

Important AI-related risk

The terms also state that Vercel may share Your Content with third-party product providers, including for training and improving their AI and machine-learning models. This creates potential confidentiality, intellectual-property, privacy, and regulatory risks, particularly where submissions contain:

• Proprietary source code or trade secrets;
• Personal data;
• Customer or employee information;
• Confidential client materials; or
• Regulated data.

The excerpt indicates that Model Training is not available or is restricted for paid Pro-plan users, but the precise scope, opt-out mechanism, exceptions, and treatment of previously collected data should be confirmed in the complete provision. The terms do not clearly establish that all customer data will be excluded from third-party providers’ training, nor do they clearly promise deletion of training data after an opt-out or termination.

2. Broader and clearer content license

The revised language reorganizes “Your Content” and expressly authorizes Vercel to create derivative works and use content for product development, security, and AI-related purposes. Although customers retain ownership, the operational license is extensive and continues to present a risk of uses beyond merely hosting or delivering the Services.

Customers represent that they own the content or have all necessary rights and permissions. This may shift responsibility to the customer for obtaining consent to use personal data or third-party materials in AI training.

3. New or emphasized AI functionality terms

The agreement now separately addresses AI Functionality and Third Party Tools. Customers are responsible for:

• Actions taken by AI features or third-party tools acting on their behalf;
• Costs incurred by those tools;
• Their configuration and use of autonomous agents; and
• Compliance with applicable laws and the Acceptable Use Policy.

This may expose customers to liability for unintended automated actions or outputs.

4. Dispute and commercial changes

The revised terms prominently incorporate:

• Binding arbitration;
• A class-action waiver;
• A 30-day arbitration opt-out;
• California governing law and San Francisco-related procedures; and
• Additional incorporated terms for AI products, APIs, marketplaces, DORA, and other services.

These provisions may materially limit litigation options and require review of multiple linked agreements.

5. Account and security responsibilities

The customer must maintain accurate account information, protect credentials, use unique logins, and accept responsibility for activities under the account. Vercel retains access rights for technical support and security purposes.

Key Changes and Risks

1. Customer Content may be used for AI model training

The revised terms add an express “Model Training” provision. By submitting content, the customer grants Vercel a broad, worldwide, royalty-free, sublicensable and transferable license to use, copy, modify, reproduce, distribute, display, publish, store, perform and create derivative works from that content.

The permitted purposes now expressly include:

• Providing and improving the Services;
• Developing new products and services;
• Securing and protecting the Services and third parties from fraud, abuse, malware and malicious content; and
• Training and improving Vercel’s artificial intelligence and machine-learning models.

The provision also states that Vercel may share customer content with third-party products, including for training and improving their AI and machine-learning models.

Risk: This is a material expansion of Vercel’s rights over customer content. The language does not appear limited to anonymized, aggregated or de-identified data, and it does not specify model-retention periods, deletion procedures, geographic limits, or whether trained models can be deleted or isolated after termination. Customers should determine whether confidential information, personal data, regulated data, source code or trade secrets may be submitted.

The excerpt indicates that paid Pro-plan customers may have a different treatment or an ability to prevent Model Training, but the surrounding text is incomplete. The exact scope, eligibility and opt-out mechanics should be confirmed before relying on that protection.

2. Broader rights to use and share customer content

The existing content license is retained but reorganized and expressly expanded to cover product development, security, fraud prevention and AI training. The license remains transferable and sublicensable, which may permit use by affiliates, contractors, technology providers and third-party AI providers.

The customer continues to bear responsibility for all submitted software, code, data, information, feedback, text and other materials, including ensuring that Vercel’s use does not infringe third-party rights or violate privacy obligations.

3. New or emphasized AI functionality risks

The terms now separately address AI functionality and third-party AI tools. If customers enable their own or third-party AI tools, agents or similar technologies to access the Services:

• Customers are responsible for resulting costs and activity;
• Those tools are treated as the customer’s electronic agents; and
• Customers may be responsible for actions taken on their behalf.

Third-party AI products and services are also subject to incorporated AI Product Terms and may have separate data-use practices.

4. Other significant changes

• The agreement more prominently incorporates arbitration and class-action waiver provisions, including a 30-day opt-out process.
• California law, San Francisco arbitration/venue provisions and individual-claim requirements are emphasized.
• Vercel retains ownership of System Data, such as telemetry, logs, traffic data and usage statistics.
• Additional terms now expressly cover AI products, APIs, DORA, Marketplace services and other offerings.

Overall assessment: The most significant commercial and privacy change is the express authorization to use and potentially share customer content for AI and machine-learning model training.

Summary

The provided diff states only that approximately 77 words were removed from the document. The actual deleted language is not included.

Legal and Commercial Impact

• Cannot determine the substance of the changes. Without the deleted wording, it is not possible to assess whether the removals affect:
• Customer data ownership or control
• Confidentiality obligations
• Data-processing permissions
• Security or privacy commitments
• Liability, indemnification, or warranties
• Termination or data-deletion rights
• Audit or compliance rights
• Intellectual-property ownership

• AI training implications cannot be evaluated. The supplied diff does not show whether language was removed concerning:
• Use of customer data to train, fine-tune, or improve AI models
• Use of customer prompts, inputs, outputs, or usage data
• De-identification, aggregation, or anonymization
• Opt-out rights or customer consent
• Retention of data for model development
• Restrictions on using customer data for third-party or general-purpose models

Potential Risk

The removal of language could either reduce or expand the provider’s rights, depending on what was deleted. In particular, deleting restrictions, consent requirements, or opt-out provisions could increase the risk that customer data may be used for AI training or product improvement without clear customer approval. Conversely, deleting a broad data-use authorization could strengthen customer protections.

Required Information

To provide a reliable legal analysis, the actual 77 deleted words—or the document sections containing them—are needed.

Summary of Changes

The supplied diff only states that approximately 77 words were removed from the document. It does not identify which clauses, sentences, or sections were deleted.

AI Training and Customer Data

• No specific changes can be assessed regarding whether customer data may be:
• Used to train, fine-tune, or improve AI models;
• Used to evaluate or test AI systems;
• Shared with AI service providers or subprocessors;
• De-identified, aggregated, or retained for model-development purposes; or
• Excluded from training by default or only upon customer request.

Because the deleted language is not provided, it is impossible to determine whether the removal:

• Expands or restricts the provider’s rights to use customer data;
• Removes a customer consent or opt-out requirement;
• Eliminates confidentiality, security, or deletion obligations;
• Changes ownership or licensing rights in customer data, inputs, outputs, or derived data; or
• Removes limitations on using personal data or regulated information for AI training.

Legal and Commercial Risk

The deletion may be material if it removed protections relating to data use, confidentiality, privacy, security, intellectual property, or AI model training. However, no reliable risk conclusion can be reached from the information provided.

Required Information

To perform a meaningful legal analysis, provide either:

1. The actual deleted text; or

2. The marked-up document showing the deleted language.

Without that text, the only confirmed change is an unspecified deletion of approximately 77 words.

Summary of Important Changes

1. Customer content may be used for AI model training

The most significant change is the express addition of a “Model Training” provision. It states that Vercel may:

• Use Your Content to train Vercel’s artificial intelligence and machine-learning models; and
• Share Your Content with third parties, including for training and improving their AI and machine-learning models.

This creates a potentially broad data-use right covering customer-submitted software, code, data, text, feedback, and other materials. The clause appears to authorize both internal training and disclosure to third-party model providers or other partners.

The diff also states: “If you are on a paid Pro plan, Model Training is not…” The remainder is not shown, so it is unclear whether Pro customers receive an opt-out, automatic exclusion, or another limitation. The final wording should be reviewed carefully. No comparable express AI-training authorization appears in the deleted language.

Risks

• Confidential code, personal information, trade secrets, or regulated data could be used in model development or disclosed to third parties.
• The clause does not, in the excerpt, clearly limit retention, anonymization, geographic processing, or deletion of training data.
• The scope of “Your Content” is broad and may include customer data embedded in applications or project files.
• Customers may need separate contractual protections, particularly for enterprise, personal, health, payment, or proprietary information.

2. Expanded content license

The customer license remains worldwide, royalty-free, transferable, and sublicensable, but the revised wording expressly permits Vercel to use content to:

• Provide and improve the Services;
• Develop new products and services;
• Secure and protect the Services; and
• Protect Vercel and third parties from fraud, abuse, malware, malicious files, and similar threats.

The separate AI-training language materially expands these rights beyond ordinary service operation.

3. AI functionality and third-party tools

The revised terms add or clarify provisions concerning AI functionality, autonomous actions, and third-party AI tools. Customers are responsible for actions taken by AI functionality or tools acting on their behalf and for costs resulting from third-party tools’ use of the Services. This increases operational and financial risk where automated agents make changes or incur usage charges.

4. Dispute-resolution changes

The revised agreement prominently adds binding arbitration and a class-action waiver, governed by California law and generally requiring individual claims. It also adds mass-arbitration procedures and a 30-day opt-out process. Customers should preserve the opt-out deadline if they do not want these provisions to apply.

5. Other material changes

• More detailed account, credential, age, and eligibility requirements.
• Broader responsibility for all activity under an account.
• Additional incorporated terms, including AI Product Terms, DORA, API, Marketplace, and other addenda.
• Continued Vercel access to accounts for technical support.
• Customer representations that submitted content may lawfully be used under the expanded license.

Key Changes and Risks

1. Customer Content may be used for AI training

Significant new provision: The revised terms expressly state that Vercel may use Your Content to train artificial intelligence and machine-learning models and may share Your Content with third-party providers, including for training and improving their AI and machine-learning models.

The new language also authorizes Vercel to use, copy, modify, reproduce, distribute, display, publish, store, perform, and create derivative works from Your Content to:

• Provide and improve the Services;
• Develop new products and services;
• Secure and protect the Services; and
• Protect Vercel and third parties from fraud, abuse, malware, and malicious content.

Customer risks

• Broad scope: “Your Content” includes code, software, data, text, inputs, feedback, and other materials submitted to or stored through the Services.
• Third-party sharing: The authorization is sublicensable and transferable, potentially allowing use by AI vendors and other service providers.
• Commercial reuse: The license is worldwide, royalty-free, fully paid, and apparently permits derivative works and product development.
• Confidentiality and privacy concerns: Customers may unintentionally submit proprietary code, personal information, regulated data, trade secrets, or customer data that could be used in model development.
• Limited apparent control: The excerpt states that, if a customer is on a paid Pro plan, “Model Training is not…”—but the remainder is missing. The full restriction, opt-out mechanism, treatment of existing training data, and availability of the restriction to other plans should be confirmed.

Recommended action: Do not submit confidential, regulated, or third-party data until the complete Model Training provision, plan-specific exclusions, deletion rights, retention rules, and AI-provider safeguards are reviewed.

2. Broader AI and third-party tool framework

A new or expanded section addresses AI Functionality and Third Party Tools. It permits AI features or connected tools to take autonomous actions and states that third-party tools operate as the customer’s “electronic agent” under UETA or similar law.

Customers bear responsibility for:

• Actions taken by connected AI tools on their behalf; and
• Costs incurred through third-party tools’ use of the Services.

This creates potential authorization, security, cost, and liability risks, particularly where AI agents can deploy code, modify projects, or access customer data.

3. Arbitration and class-action waiver made prominent

The revised presentation prominently requires review and acknowledgment of mutual binding arbitration and a class-action waiver. The terms require individual arbitration, generally waive class and representative proceedings, and provide a 30-day opt-out process. Customers should calendar the opt-out deadline if they do not accept these restrictions.

4. Other notable changes

• New or clearer age, device-control, account-information, unique-login, and account-activity requirements.
• Open-source software, including Next.js, is expressly excluded from the Agreement.
• Additional incorporated terms are highlighted, including AI Product Terms, DORA, API, Marketplace, and other addenda.
• Vercel retains broad rights to access accounts for technical support and to use customer feedback.
• The customer’s license to Vercel is expressly sublicensable and transferable, increasing downstream-use risk.

Summary of Change

• The document removes approximately 77 words.
• The actual deleted wording is not provided, so the legal and operational effect of the change cannot be determined reliably.

AI Training and Data Use

• The diff does not identify whether the deleted text addressed:
• Use of customer data to train, fine-tune, or improve AI models;
• Whether customer data may be used in anonymized, aggregated, or de-identified form;
• Restrictions on using prompts, inputs, outputs, or telemetry for model development;
• Opt-out or consent requirements;
• Ownership or licensing rights in customer data or AI outputs; or
• Human review, data retention, or disclosure to AI service providers.
• If the removed language contained a prohibition or limitation on AI training, its deletion could materially expand the provider’s ability to use customer data.
• Conversely, if the deleted language granted training rights or broad data-use permissions, its deletion could restrict those rights or create uncertainty.

Key Legal Risks

• Interpretation risk: Without the deleted text, it is impossible to determine whether the change narrows or expands either party’s rights.
• Data-protection risk: Removal of privacy or data-use limitations could affect compliance with applicable data-protection laws and customer commitments.
• Contractual ambiguity: If related definitions, exceptions, or consent mechanisms remain elsewhere in the agreement, deleting only part of a provision could create inconsistencies.
• Evidence and negotiation risk: The parties may later disagree about whether the deletion was intended to change substantive rights or merely simplify wording.

Recommended Review

Obtain the complete before-and-after versions, or at least the exact 77 deleted words, and compare them with provisions concerning:

1. Customer data ownership and permitted uses;

2. AI model training and service improvement;

3. De-identification and aggregated data;

4. Data retention, deletion, and subprocessors;

5. Confidentiality and privacy obligations; and

6. Customer opt-out or consent rights.

No definitive conclusion about AI-training rights should be drawn from the word-count description alone.

Summary

The diff only states that approximately 77 words were removed from the document. It does not identify the deleted language.

Key Legal and Commercial Implications

• Unable to assess substantive changes: Without the actual deleted text, it is not possible to determine whether the changes affect liability, confidentiality, intellectual property, termination, data protection, or other contractual rights.
• AI training provisions: The diff provides no information about whether customer data may now be used—or is no longer permitted to be used—to train, fine-tune, evaluate, or improve AI models.
• Potential risk from omitted protections: If the removed language concerned restrictions on data use, confidentiality, security, ownership, deletion, or opt-out rights, its deletion could materially increase the customer’s exposure.
• Potential benefit from omitted permissions: Conversely, the deletion may have removed a provider’s authorization to use customer data for model training or product improvement. This cannot be confirmed from the available information.

Required Information

Please provide the actual 77-word deletion, preferably with surrounding text or the full before-and-after clause. That is necessary to determine:

1. Whether customer data can be used to train or improve AI models;

2. Whether de-identified, aggregated, or derived data remains usable;

3. Whether the provider must obtain consent or offer an opt-out;

4. Whether data-use rights survive termination; and

5. Whether the change affects confidentiality, ownership, security, or regulatory compliance.