Sketch

Summary of Important Changes

1. Major restructuring and incorporation of data-processing documents

• The document is no longer presented simply as an “End User License Agreement.” It now references a Data Processing Addendum, Annex 1 – Details of Processing, and Annex 2 – Security Measures.
• The revised navigation also refers to subprocessors, privacy statements, security measures, and related policies.
• These documents may now form part of the contractual framework or be incorporated by reference. Their exact contractual status, priority over the EULA, and ability to be amended should be confirmed.

Risk: Important privacy, security, and processing obligations may be located outside the main EULA, potentially making the customer responsible for reviewing multiple linked documents.

2. Expanded data collection and use

The revised “Use of data” clause states that Sketch and its subsidiaries may periodically collect and use:

• Technical and related data concerning the Sketch Mac App;
• The licensed app’s version number;
• Information about the system on which the app is installed.

Sketch may use this information to:

• Facilitate maintenance and support;
• Improve its products;
• Provide further services and technologies.

The clause also states that Sketch may process personal data and will comply with applicable data-protection law, directing users to the privacy statement for further details.

Risk: The permitted purposes are broader than simple diagnostics or support. “Improve its products” and “provide further services and technologies” are open-ended and may allow extensive analytics or product-development uses. The clause does not specify data minimization, retention periods, aggregation/anonymization standards, or whether data may be shared with affiliates or subprocessors.

3. AI-model training

• The diff does not expressly state that customer data, customer content, usage data, telemetry, or personal data may be used to train, fine-tune, evaluate, or improve AI or machine-learning models.
• It also does not expressly prohibit such use.
• References to “LLMs,” “Data Processing,” and “Details of Processing” appear primarily to concern document presentation and privacy documentation, not a clear grant or restriction relating to AI training.
• The broad purposes—particularly “improve its products” and “provide further services and technologies”—could create ambiguity if Sketch later uses telemetry or other data in AI development.

Recommended clarification: Require an express statement addressing whether customer content, prompts, designs, telemetry, metadata, and personal data may be used for AI training; whether use is opt-in or opt-out; whether data is anonymized; whether third-party model providers receive it; and whether customer data is excluded from model training by default.

4. Liability changes

• The revised language generally caps liability at amounts paid to Sketch in the preceding 12 months.
• The previous exceptions for negligence causing death or physical injury, and for intentional misconduct or gross negligence, appear to be retained in modified form.
• The revised wording may be less clear and appears to reorganize or narrow the exclusions from the liability cap.

Risk: Recovery for data loss, confidentiality breaches, privacy violations, or unauthorized AI use may remain subject to the general cap unless expressly carved out.

5. Other contractual changes

• Corporate users must warrant that the accepting individual has authority to bind the entity.
• Definitions and company-identification details have been updated.
• Documentation and system requirements are now tied to Sketch’s website and may change over time.
• The warranty disclaimer and “as is/as available” language remains prominent.
• The EULA continues until expiry or termination of the Sketch Mac License.

Summary

Key limitation

The supplied diff does not include the actual amended contract language. It only states:

> “Added approximately 294 words to the document”

Without the text of those additions—and any surrounding provisions affected—it is not possible to reliably identify:

• New legal obligations or rights
• Changes to liability, confidentiality, ownership, or termination
• Whether customer data may be used to train AI models
• Whether any AI-training permission is opt-in, opt-out, limited, or perpetual
• Whether data is anonymized, aggregated, or shared with third parties
• Whether the customer’s data may be used to improve products or services
• Whether customer content may be reviewed by humans or processed by subcontractors

AI-training review

No conclusion can be reached about AI-model training from the material provided. The diff does not contain any language addressing:

• Training, fine-tuning, or improving artificial intelligence or machine-learning models
• Use of prompts, inputs, outputs, customer content, or usage data for model development
• De-identification, aggregation, or retention of data
• Customer consent or the ability to opt out
• Ownership or licensing of data and model outputs

Information needed

Please provide the actual 294 words added to the contract, together with any deleted or replaced wording shown using the stated notation:

• Additions: {new text}
• Deletions: [old text]
• Replacements: [old text]{new text}

The surrounding section headings and existing language would also be useful, because the legal effect of an addition often depends on how it interacts with confidentiality, data-processing, intellectual-property, and limitation-of-liability provisions.

Summary

The provided diff does not include the actual amended legal language. It only states:

> “Added approximately 294 words to the document”

Accordingly, it is not possible to determine:

• Which provisions were added or changed;
• Whether customer data may be used to train, fine-tune, or improve AI models;
• Whether data may be shared with affiliates, service providers, or third parties;
• Whether the customer’s data may be retained after termination;
• Whether the provider receives ownership or broad usage rights;
• Whether confidential information or personal data is subject to new processing terms;
• Whether any consent, opt-out, deletion, security, or audit rights have changed; or
• Whether liability, indemnity, or regulatory obligations have been expanded.

AI-Training Risk Assessment

No substantive assessment can be made because the added wording is missing. In particular, the diff should be reviewed for terms such as:

• “train,” “fine-tune,” “improve,” or “develop” models or services;
• “inputs,” “outputs,” “customer data,” or “usage data”;
• rights that are “worldwide,” “perpetual,” “irrevocable,” “royalty-free,” or transferable;
• permission to use aggregated, de-identified, or anonymized data;
• sharing data with model providers, affiliates, or subcontractors;
• retention of data after account termination; and
• opt-out, deletion, confidentiality, and security limitations.

Required Information

Please provide the actual text of the diff, including the approximately 294 added words and any surrounding deleted or replacement language. Without the substantive wording, the legal and AI-data-use implications cannot be reliably analyzed.

Summary of Important Changes

1. New data-processing materials and annexes

The EULA is being reorganized to include or link to:

• A Data Processing Addendum
• Annex 1 – Details of Processing
• Annex 2 – Security Measures
• Subprocessor information and a Privacy Statement

These materials may become contractually relevant, but the diff does not clearly state their order of precedence or whether they apply to all customers. Customers should confirm that the linked documents are available, complete, and consistent with the EULA.

2. Expanded collection and use of technical data

The new wording expressly permits Sketch and its subsidiaries to collect and use “technical and related data” concerning:

• The version of the Sketch Mac App installed
• The system on which it is installed

Sketch may use this data to:

• Maintain and support the App
• Improve its products
• Provide further services and technologies

This is broader than the previous warranty disclaimer, which primarily stated that the App was provided “as is.” The new language creates an express contractual basis for telemetry and product-improvement processing.

3. Personal-data processing

The revised EULA states that Sketch may process personal data where necessary to:

• Provide the License Key
• Maintain and support the App
• Comply with legal obligations

It refers customers to the Privacy Statement for further details regarding personal data collected through the website. The EULA does not itself specify categories of personal data, retention periods, international transfers, customer rights, or detailed security obligations; those matters appear to be moved to the Privacy Statement and Data Processing Addendum.

4. AI-model training: no express authorization identified

The diff contains references to “LLMs” and “Copy for LLMs,” but these appear to describe document-viewing or formatting tools rather than substantive customer-data rights.

There is no clear clause expressly stating that customer content, Workspace data, designs, prompts, outputs, or other Customer Data may be used to train, fine-tune, evaluate, or improve AI or machine-learning models.

However, the new permission to use technical and related data to “improve its products” and provide “further services or technologies” is relatively broad. Depending on the definitions in the missing annexes or Privacy Statement, Sketch could potentially argue that certain telemetry or usage data may support AI-related product development. Customers should seek written confirmation that Customer Data and content are excluded from model training unless expressly authorized.

5. Corporate authority and acceptance

The revised acceptance language expressly requires a person accepting for a company or legal entity to warrant that they have authority to bind it. It also clarifies that minors may not accept the EULA.

6. Liability and warranty restructuring

The liability provisions have been substantially redrafted but retain a cap generally tied to amounts paid during the preceding 12 months. The App remains provided on an “as is” and “as available” basis, with broad exclusions for consequential and indirect losses, including loss of data. Customers should assess whether the revised wording narrows available remedies for data-related incidents.

Key Changes and Risks

1. AI processing and model training

• Third-party AI providers may use data for model training. The revised statement says that requests from Sketch may be sent to a user-selected third-party AI provider. That provider’s privacy policy governs how it processes, retains, and “may use this data, including for model training.”
• Risk: Customer content, prompts, images, or related personal data may be retained or used to train external models, depending on the provider’s terms and settings.
• The statement does not specify whether training is opt-in, whether data is anonymized, how long it is retained, or whether customers can require deletion or prohibit training.
• Customers are told to review the provider’s privacy policy, tool permissions, and settings. This shifts much of the practical risk and responsibility to the customer.

• Sketch’s own use of data for training remains broad. The revised purposes section continues to permit processing personal data to “improve” services and for “training and quality control purposes.”
• Risk: The language does not clearly limit training to fraud prevention, service functionality, or anonymized data. It could potentially cover product or machine-learning development.
• The statement does not expressly distinguish between account/support data, usage data, uploaded design content, prompts, or other customer-submitted materials.

• Local processing protections are clarified.
• Background removal is described as operating locally on the Mac using Apple’s on-device machine-learning frameworks, with images not sent to Sketch or an external provider for that processing.
• Sketch’s Model Context Protocol (MCP) server runs locally and can be disabled in settings.
• However, requests or information sent through connected third-party AI tools may leave the device and be processed under the third party’s terms.
• These protections therefore apply only to specified features and do not prevent external processing through integrations.

2. Expanded or clarified data collection

The revised statement expressly covers the website, Mac app, Web app, and iOS app and lists potentially extensive data, including:

• Account, contact, payment, support, discussion, and research information.
• Student-status documents that may include date of birth.
• IP address and estimated country.
• Device identifiers and technical information, including Mac UDID, hardware details, operating-system version, timestamps, and license identifiers.
• Information supplied by an employer, reseller, agent, or other customer organization.

Risk: Collection is more specific but remains broad, including a residual right to collect other data when necessary for stated purposes.

3. Legal bases and disclosures

• The revised statement reorganizes legal bases around contractual necessity, legal obligations, legitimate interests, and consent.
• “Legitimate interests” expressly include improving products and services, preventing fraud, and reporting possible criminal conduct.
• Data may be shared with service providers, hosting and IT providers, payment and communications providers, subprocessors, authorities, and business transferees.
• A new privacy contact email, privacy@sketch.com, is provided.

Overall assessment

The most significant customer risk is that AI integrations can transfer customer data to external providers that may retain or use it for model training. The statement provides no uniform no-training commitment, retention limit, or customer-controlled opt-out for such providers.

Summary

Key limitation

The provided diff does not include the text of the additions, deletions, or replacements. It only states:

> “Added approximately 338 words to the document”

Without the actual amended language, it is not possible to reliably identify:

• New legal obligations or liabilities
• Changes to confidentiality, security, or data-retention terms
• Changes to intellectual-property ownership or licensing
• Changes to liability limits, indemnities, or warranties
• Changes to termination, governing law, or dispute-resolution provisions
• Any provisions concerning the use of customer data to train, improve, or evaluate AI models

AI-training implications

No conclusion can be drawn about AI-model training from the information provided. The added text could potentially:

• Permit use of customer data to train or fine-tune models
• Permit use of data to improve products or services, which may indirectly include model training
• Allow use of aggregated, anonymized, or de-identified data
• Restrict training use to customer-directed or opt-in circumstances
• Prohibit use of customer content for training
• Reserve broad rights to use prompts, inputs, outputs, metadata, or telemetry

The relevant language should be reviewed for terms such as “train,” “training,” “fine-tune,” “machine learning,” “artificial intelligence,” “improve,” “develop,” “aggregate,” “de-identify,” “customer content,” “inputs,” “outputs,” and “service data.”

Information needed

Please provide the full redline or the actual added, deleted, and replacement text. Once supplied, the review can identify:

1. The precise contractual change

2. Its practical and legal effect

3. Any customer-side risks

4. Whether customer data may be used for AI training

5. Whether consent, opt-out, anonymization, retention, or deletion protections apply

Summary of Important Changes

1. AI and data use

• New local-processing disclosures: Background removal is stated to use Apple’s on-device machine-learning frameworks. Images are not uploaded to Sketch or third-party servers for that processing. This reduces cloud-transfer risk, but the provision appears limited specifically to background removal and should not be read as a general promise that all AI features operate locally.

• New Model Context Protocol (“MCP”) feature: Sketch’s MCP server runs locally on the customer’s Mac and is disabled by default. The customer chooses whether to enable it and which compatible third-party AI tools may connect.

• Broad access and action rights for connected AI tools: Connected third-party AI tools may request access to document information and may perform actions including:
• modifying documents; and
• exporting Content.

The customer can disable the MCP server in Sketch settings, but approval controls depend on the connected AI tool and its configuration. This creates a risk that a connected tool may access, alter, or export sensitive Workspace information, potentially with limited control from Sketch.

• Disclosure to third-party AI providers: Information obtained from Sketch may be transmitted to the customer’s chosen third-party AI provider. That provider’s terms and privacy policy govern its processing in addition to Sketch’s Terms.

• Model-training and retention settings: Customers are expressly told to review the third-party AI provider’s permissions and settings, including data retention and model-training settings, before connecting the tool. This is an important allocation of risk: the Terms do not prohibit the provider from retaining customer data or using it to train AI models. The customer is responsible for checking whether training is enabled and adjusting the provider’s settings.

• No express Sketch training right identified: The new language does not itself grant Sketch a right to use customer Content to train Sketch’s AI models. However, Content transmitted to a third-party provider may be used for model training if that provider’s terms or settings permit it. The provision also does not clearly state whether Sketch has contractual responsibility for the provider’s handling of the data.

2. Customer control and security risks

• MCP is opt-in and can be disabled, which is a helpful safeguard.
• However, permissions depend on the connected AI tool, and the Terms do not promise granular approval, read-only access, confirmation before exports, or prevention of sensitive-data transmission.
• Customers should assess provider terms, training defaults, retention periods, human review, security practices, and deletion controls before enabling MCP.

3. Contractual and administrative changes

• The Terms date changes from 22 April 2026 to 06 October 2026.
• Customer assignment or subcontracting rights remain restricted and now expressly cover assignment of rights or obligations to third parties.
• Sketch may assign or subcontract its rights and obligations at its sole discretion.
• Sketch’s permitted assignees and successors may enforce the Terms.
• The “entire agreement” and third-party-terms language is reorganized and expanded. Purchase orders, portals, codes of conduct, and similar customer documentation generally do not override the Terms unless Sketch expressly references and waives its rights.
• Dutch Civil Code article 6:254 is expressly excluded for third-party stipulations.
• Failure to enforce a provision is not a waiver; invalid provisions do not affect the remainder of the Agreement.

Summary

The provided diff does not include the actual amended contract language. It only states:

> “Added approximately 338 words to the document”

Accordingly, it is not possible to determine:

• What contractual terms were added or changed;
• Whether customer data may be used to train, fine-tune, or improve AI models;
• Whether any consent, opt-out, or deletion rights were added or removed;
• Whether data may be shared with affiliates, vendors, or third parties;
• Whether confidentiality, security, ownership, or liability provisions changed; or
• Whether the customer assumes additional legal or financial risk.

AI Training and Data-Use Review

No conclusions can be drawn about AI-model training because the actual additions are not provided. The relevant language would typically address whether the provider may:

• Use customer content, prompts, outputs, or usage data to train or improve models;
• Use data in aggregated, de-identified, or identifiable form;
• Retain data after termination;
• Permit human review or use data for safety monitoring;
• Share data with subprocessors or model providers;
• Offer an opt-out or require affirmative consent; or
• Exclude customer data from training by default.

Information Needed

Please provide the full diff, including the text marked as:

• {additions};
• [deletions]; and
• []{replacements}.

Without the underlying wording, any assessment of legal effect or risk would be speculative.

Diff Analysis

Key Limitation

The provided diff does not include the actual contractual language. It only states:

> “Added approximately 294 words to the document”

Without the text of the additions, deletions, or replacements, it is not possible to determine:

• What legal terms changed;
• Whether obligations or liability shifted;
• Whether customer rights were reduced;
• Whether new data-sharing or confidentiality provisions were added; or
• Whether the customer’s data may be used to train artificial-intelligence (AI) models.

AI Training and Data Use

No conclusion can be drawn about AI-model training from the information provided. The diff should be reviewed specifically for language addressing:

• Use of customer content, data, prompts, inputs, outputs, or usage information;
• Training, fine-tuning, improving, testing, or evaluating AI or machine-learning models;
• Whether data is used for product development or service improvement;
• Whether use is automatic, optional, or subject to customer consent;
• Opt-out or deletion rights;
• Use of data in aggregated, de-identified, or anonymized form;
• Sharing with affiliates, vendors, subprocessors, or model providers;
• Retention periods and deletion obligations;
• Ownership of customer data and AI-generated outputs; and
• Confidentiality, security, and restrictions on human review.

Information Needed

Please provide the full redlined language, including the words shown in {additions}, [deletions], and []{replacements}. Once provided, the changes can be analyzed for:

1. New or expanded provider rights;

2. New customer obligations;

3. Changes to confidentiality, privacy, or security protections;

4. Liability, indemnity, warranty, and termination effects; and

5. Any permission to use customer data to train or improve AI models.

Summary

The provided diff does not include the actual amended contractual language. It only states:

> “Added approximately 191 words to the document”

Accordingly, it is not possible to determine:

• What provisions were added, deleted, or replaced;
• Whether the customer’s data may be used to train, fine-tune, evaluate, or improve AI models;
• Whether customer data may be shared with affiliates, vendors, or model providers;
• Whether data is anonymized, aggregated, de-identified, or retained after termination;
• Whether the customer has any opt-out, deletion, or control rights;
• Whether confidentiality, security, intellectual-property, or indemnity obligations changed; or
• Whether the provider receives broader rights to use customer content or inputs.

AI Training Risk Assessment

No conclusion can be reached from the supplied material. The diff must include the specific added or modified wording to assess whether it:

• Authorizes training on customer prompts, inputs, outputs, files, or usage data;
• Permits use for general model training rather than only providing the contracted service;
• Applies by default unless the customer opts out;
• Allows human review or use by third-party service providers;
• Grants the provider rights to retain or reuse data after account termination; or
• Limits the provider’s responsibility for models trained using customer data.

Information Needed

Please provide the actual 191-word addition and any surrounding language, using the stated notation:

• {added text}
• [deleted text]
• []{replacement text}

Without the contractual text, there are no identifiable legal changes or specific risks to analyze.

Key Changes and Risks

1. AI and training-related data use

• The revised statement continues to permit Sketch to record and monitor communications with users.
• The stated purposes now include maintaining records, checking instructions, analysing and improving services, and “training and quality control purposes.”
• The earlier wording referred more generally to using data for “training and quality control purposes”; the revised wording places this permission within the section on support communications and legitimate interests.
• No express reference to AI, machine learning, foundation models, model training, or use of customer content to train AI models appears in the diff. Accordingly, the statement does not clearly confirm whether:
• customer designs, files, prompts, or other platform content may be used to train AI models;
• communications may be supplied to AI vendors;
• training data is anonymised or de-identified;
• customers can opt out; or
• trained models retain or reproduce customer information.

Risk: “Training” is broad and could potentially be interpreted to include AI-system training, particularly when combined with “analyse and improve our services.” The statement does not define the scope, data categories, retention period, model providers, safeguards, or customer controls. This creates uncertainty and potential GDPR purpose-limitation/transparency concerns, especially for confidential business or design data.

2. Expanded and reorganised data collection

The revised statement more clearly identifies Sketch B.V. as the controller and expands the description of data collected, including:

• contact and account information;
• payment details;
• student-status evidence, potentially including date of birth;
• IP address and country;
• browser, computer/device information, timestamps, license identifiers, and Mac UDID;
• information from employers or other customer organisations.

The Mac and iOS app data descriptions are more detailed than before, and the policy now states that additional personal data may be collected where necessary for the listed purposes.

Risk: The broader “other types of personal data” language may permit collection beyond the enumerated categories without clear limits.

3. Revised legal bases and purposes

The revised text expressly lists:

• contractual necessity;
• legal obligations;
• legitimate interests, including improving products, fraud prevention, and reporting criminal acts; and
• consent for specified processing.

It also states that legitimate interests will be used only where reasonable and balanced against users’ rights and freedoms.

Risk: The legitimate-interest basis may be relied on for service improvement and communications monitoring, potentially reducing the need for separate consent. The relationship between consent-based “training” and legitimate-interest-based service improvement is unclear.

4. Sharing and transfers

The policy now describes sharing with service providers, IT and hosting providers, payment and communications providers, authorities, and potential business transferees. It refers to the Subprocessors list and Standard Contractual Clauses for international transfers.

Risk: A sale or transfer of the business may transfer user data to an acquiring entity, with only a general commitment to consistent use.

5. User rights and contact details

The revised statement adds privacy@sketch.com and clearer language about updating data and exercising rights, while retaining the right to complain to a supervisory authority.

Structured Summary of Important Changes

1. Data use and AI-model training

• The revised EULA adds a new “Use of data” provision. Sketch and its subsidiaries may collect and use technical and related data concerning the licensed Sketch Mac App, including:
• The version of the App installed; and
• The system on which it is installed.
• Sketch may use this data to:
• Maintain and support the App;
• Improve its products; and
• Provide further services or technologies.
• Sketch may also process personal data where necessary to provide the License Key, support and maintenance, or comply with applicable data-protection law. The EULA refers users to Sketch’s Privacy Statement for more details.

AI-specific risk

• The diff does not expressly state that customer data, designs, content, prompts, outputs, telemetry or other information may be used to train, fine-tune or evaluate AI models.
• However, the new wording permits use of broadly defined “technical and related data” to improve products and provide technologies. Depending on the Privacy Statement or incorporated policies, this could potentially include data used in machine-learning or AI development.
• The EULA does not clearly:
• Exclude customer-created content or design files from AI training;
• Prohibit use of customer data to train third-party or general-purpose AI models;
• Explain whether data is anonymized or aggregated;
• Provide an opt-out from AI-related processing; or
• State whether customer data is retained after account or license termination.
• Customers should review the referenced Privacy Statement, Data Processing Addendum and any LLM/AI annexes to determine whether broader AI-processing rights exist outside the EULA.

2. Privacy and data-protection framework

• The EULA now expressly states that Sketch will comply with applicable data-protection law.
• More processing detail is moved or deferred to the Privacy Statement, creating a risk that important data-use terms are located outside the EULA and may be changed separately.
• The new language appears focused primarily on Sketch’s own technical data and personal data, rather than expressly addressing customer-uploaded content or confidential business information.

3. Acceptance and authority

• Acceptance mechanics are reorganized and expanded. A person accepting for a company warrants that they have authority to bind it.
• Users must be of legal age and may not accept the EULA if they cannot form a binding contract.
• Agreement is still formed through purchasing, downloading or using the App.

4. Liability and warranties

• The revised text preserves or restates broad “as is” and “as available” disclaimers.
• Liability remains generally capped at amounts paid to Sketch in the preceding 12 months, while excluding consequential or indirect damages, including loss of data.
• The revised structure may make the cap and exclusions apply more broadly to Sketch and third parties involved in creating, producing or delivering the App.

5. Other changes

• Definitions are reorganized and clarified, including Sketch’s legal identity, License Key, Open Source Software, Documentation, Updates and Use.
• Termination language is clarified: the EULA continues until expiry or termination of the Sketch Mac License, with automatic termination in specified circumstances.

Summary

Information Provided

The diff only states:

> “Added approximately 294 words to the document”

No actual added, deleted, or replaced contract language is included.

AI Training and Customer Data

I cannot determine whether the changes:

• Permit the provider to use customer data, prompts, outputs, or other content to train AI models;
• Restrict or prohibit such training;
• Allow use of data in anonymized, aggregated, or de-identified form;
• Permit human review or use by vendors for model improvement;
• Grant the provider rights to retain data after termination; or
• Create exceptions for abuse monitoring, safety testing, or service improvement.

These issues depend on the precise wording of the added provisions.

Other Legal Risks

The available diff is also insufficient to assess changes involving:

• Data ownership and intellectual-property rights;
• Confidentiality obligations;
• Privacy-law compliance;
• Data retention and deletion;
• Security commitments and breach liability;
• Subprocessors or third-party service providers;
• Customer indemnities;
• Limitation of liability;
• Warranty disclaimers;
• Audit rights; or
• Termination consequences.

Required Information

Please provide the actual 294-word addition and any surrounding text, using the stated notation:

• {new text}
• [deleted text]
• [old text]{new text}

Without the underlying language, no reliable legal comparison or identification of AI-training risks is possible.

Structured Summary of Important Changes

1. New data-processing materials and AI references

• The document is now presented with links or references to a Data Processing Addendum, Processing Annexes, Security Measures, Subprocessors, and a Privacy Statement.
• The navigation/title includes references to “Details of Processing,” “LLMs,” and “Copy for LLMs.” These appear primarily to be document-formatting or navigation changes, not operative contractual language.
• The diff does not expressly state that customer content, designs, files, prompts, or other customer data may be used to train AI models.
• It also does not expressly prohibit such training. The operative data-use language is therefore potentially incomplete or ambiguous from an AI-governance perspective. The referenced Privacy Statement, Data Processing Addendum, and Annexes should be reviewed because they may contain the controlling provisions.

2. Expanded technical-data collection and use

The former general warranty disclaimer is replaced with a detailed data-use clause permitting Sketch and its subsidiaries to periodically collect and use:

• Technical and related data concerning the licensed Sketch Mac App;
• The version of the application installed; and
• Information about the system on which it is installed.

Sketch may use this data to:

• Maintain and support the application;
• Improve its products;
• Provide further services and technologies; and
• Comply with legal obligations.

This is a material expansion and gives Sketch broader operational and product-improvement rights. It does not expressly authorize using customer content to train AI models, but “improve its products” and “provide further services and technologies” could be broad enough to warrant clarification.

3. Personal-data processing

• Sketch may process personal data where necessary to provide the License Key, maintenance, support, and comply with applicable data-protection law.
• The EULA directs users to the Privacy Statement for more details concerning personal data collected through the website.
• The operative language is less self-contained than before and incorporates external policies. Customers should verify whether those documents permit analytics, model development, profiling, or sharing with subprocessors.

4. Acceptance and authority

• Acceptance is now expressly tied to purchasing a License Key and/or downloading or using the Mac App.
• Representatives accepting for an entity warrant that they have authority to bind it.
• Minors may not accept the EULA if they lack legal capacity.

5. Definitions and licensing

• Definitions are reorganized and clarified, including Sketch, the Sketch Mac App, License Key, Open Source Software, Documentation, Use, and Updates.
• Documentation is now tied to Sketch’s website and may be amended over time.
• The License Key remains a code enabling activation and use of the App.

6. Liability and warranty changes

• The liability cap is restated as amounts paid during the preceding 12 months.
• Broad exclusions for consequential, indirect, lost-profit, business-interruption, and data-loss damages remain.
• Exceptions are preserved for death or physical injury caused by negligence and damages arising from intent, willful misconduct, or gross negligence.
• The App and documentation remain provided “as is” and “as available,” with no fitness warranty.

Key action: Obtain and review the referenced Privacy Statement, Data Processing Addendum, processing annexes, and subprocessor list specifically for any express or implied AI-training rights.

Summary

The provided diff does not include the actual added, deleted, or replaced contractual language. It only states:

> “Added approximately 294 words to the document”

Accordingly, it is not possible to determine:

• What provisions changed;
• Whether the changes expand or restrict either party’s rights;
• Whether liability, confidentiality, intellectual property, data protection, or termination terms were modified; or
• Whether the customer’s data may be used to train, fine-tune, evaluate, or improve AI models.

AI-Training and Data-Use Risks

No specific AI-related language is visible in the supplied diff. The following issues should be checked in the missing text:

• Training permission: Whether the provider receives an express or implied right to use customer data to train or fine-tune artificial intelligence or machine-learning models.
• Scope of data: Whether “customer data” includes prompts, inputs, outputs, uploaded files, personal data, confidential information, or usage telemetry.
• De-identification: Whether data may be anonymized, aggregated, or pseudonymized before use, and whether the provider must actually remove identifying information.
• Purpose limitations: Whether data may be used only to provide the services or also to improve products, develop models, conduct research, or create commercial products.
• Opt-out rights: Whether the customer can opt out of training or model-improvement uses, and whether the opt-out applies retroactively.
• Subprocessors and affiliates: Whether data may be shared with affiliates, contractors, cloud providers, or third-party model developers for AI-related purposes.
• Retention and deletion: Whether training copies, derived datasets, model weights, embeddings, or logs must be deleted when the agreement ends.
• Ownership and confidentiality: Whether the customer retains ownership and whether trained models or outputs could incorporate or disclose customer information.
• Regulatory compliance: Whether the provider assumes responsibility for privacy, confidentiality, sector-specific, or cross-border data-transfer requirements.

Required Information

Please provide the full redline or the actual 294 words of added text, including any surrounding provisions needed to interpret them. Without that language, no reliable legal-risk assessment or comparison can be made.

Summary

The provided diff does not include the actual amended legal language. It only states:

> “Added approximately 338 words to the document”

Because the additions, deletions, and replacements are not shown, it is not possible to reliably determine:

• What contractual obligations changed;
• Whether liability, indemnity, confidentiality, or termination rights were modified;
• Whether customer data may be used for AI or machine-learning training;
• Whether the customer’s consent is required for such use;
• Whether data may be shared with affiliates, vendors, or third parties;
• Whether customer data may be retained after termination; or
• Whether the customer received any new audit, opt-out, deletion, or security rights.

AI-Training Review

No conclusions can be drawn about AI-model training from the information provided. The added 338 words could potentially address AI training, but the actual text is necessary to assess:

• Whether customer data, prompts, outputs, or usage data may be used to train or improve models;
• Whether use is limited to de-identified or aggregated data;
• Whether data may be used for the provider’s general-purpose models or only for the customer’s account;
• Whether human review or third-party processing is permitted;
• Whether the customer can opt out; and
• Whether the provider must delete or stop using data upon request or contract termination.

Required for Further Analysis

Please provide the actual redlined text using the stated notation, for example:

• Additions: {new language}
• Deletions: [deleted language]
• Replacements: [old language]{new language}

Without the underlying language, any assessment of legal effect or AI-data-use risk would be speculative.

Structured Summary of Important Changes

1. New data-processing materials and cross-references

• The EULA is now presented alongside or linked to a Data Processing Addendum, Annex 1 (“Details of Processing”), Annex 2 (“Security Measures”), Subprocessors information, Privacy Statement, and related policies.
• These materials may contain important obligations and processing details that are not included in the EULA itself. Customers should review whether they are incorporated by reference and which document controls in the event of conflict.
• The revised text identifies Sketch B.V. with its Dutch registration and VAT details.

2. Customer and organizational acceptance

• The agreement now expressly permits acceptance on behalf of a company or other legal entity, but requires the individual accepting it to warrant that they have authority to bind that entity.
• It also states that a person may not accept the EULA if they are not legally old enough to form a binding contract.
• This creates potential exposure for unauthorized representatives or organizations if the authority warranty is inaccurate.

3. Use and collection of technical data

The revised “Use of data” language allows Sketch and its subsidiaries to periodically collect and use:

• The version number of the Sketch Mac App installed;
• Information about the system on which it is installed; and
• Other “technical and related data concerning” the licensed application.

The stated purposes are to:

• Facilitate maintenance and support;
• Improve Sketch products;
• Provide further services and technologies; and
• Fulfil obligations under the EULA.

Sketch also states that it may process personal data and directs users to its Privacy Statement for further details. This is broader and more operationally specific than the previous warranty-focused wording.

4. AI model training

• No express authorization to use customer data to train AI models appears in the supplied diff.
• The revised purposes refer to improving products and providing further services or technologies, but do not expressly mention:
• Training, fine-tuning, or evaluating AI or machine-learning models;
• Generative AI development;
• Using customer content, designs, prompts, or outputs as training data; or
• De-identification, aggregation, retention, or model-output ownership.
• Accordingly, the diff does not clearly create a contractual right to train AI models on customer data. However, the broad “improve products” and “provide further services or technologies” wording could create ambiguity, particularly when read with the Privacy Statement or Data Processing Addendum. Customers should confirm those documents contain an explicit prohibition or permission regarding AI training.

5. Liability and warranty changes

• The liability provision is substantially reorganized and now generally caps liability at amounts paid to Sketch in the preceding 12 months.
• Consequential and indirect damages—including loss of profit, business interruption, and loss of data—remain broadly excluded.
• Exceptions are retained for death or physical injury caused by negligence and damage arising from intent, wilful misconduct, or gross negligence.
• The EULA continues to provide the App and documentation on an “as is” and “as available” basis, with broad warranty disclaimers.

Summary of Important Changes

1. AI-model and training-related use of data

• The revised statement expressly permits Sketch to record and monitor communications between the customer and Sketch.
• It states that communications may be used to:
• Maintain records;
• Check customer instructions;
• Analyze, assess, and improve services; and
• Conduct “training and quality control purposes.”
• The prior wording referred more narrowly to recording or monitoring communications for support, important updates, or newsletter purposes. The new wording creates a broader operational and improvement use.
• The statement does not clearly say that customer designs, project files, or other Platform content will be used to train general-purpose AI models. It also does not define:
• What “training” means;
• Whether training involves machine-learning or AI models;
• Whether data is anonymized or aggregated;
• Whether human reviewers or third-party AI providers are involved;
• How long training data or model outputs are retained; or
• Whether customers can opt out.
• The new language identifies consent as a legal basis for processing data for a “specific purpose,” but it is unclear whether separate, affirmative consent is required for training, whether consent is optional, or how consent can be withdrawn. This ambiguity is a material risk, particularly for confidential business communications and enterprise customers.

2. Broader and more detailed data collection

The revised statement expands and reorganizes the categories of data collected, including:

• Contact details, including name, address, email, telephone number, and spoken language;
• Account information, including password, profile image, username, alias, and license key;
• Payment information, including cardholder name, last four digits, expiry date, card type, issuing bank, and country of origin;
• Proof of student status, potentially including date of birth;
• IP address and inferred country;
• Device identifiers and technical details, including computer name, model, GPU/CPU type, UDID, operating-system version, timestamps, and license identifier; and
• Records of support and other discussions.

Sketch also reserves the right to collect other personal data where required by law or reasonably necessary for the purposes in the statement.

3. New or expanded purposes and legal bases

• “Legitimate interests” are described more broadly to include improving products and services, preventing fraud, and reporting possible criminal acts.
• Processing for service improvement, training, and quality control is now more explicit.
• Marketing and analytics purposes are reorganized, with consent required where legally applicable and opt-out rights for marketing communications.

4. Sharing and transfers

• The statement more clearly identifies service providers, IT/data-hosting providers, payment providers, and communications providers as recipients.
• It adds or clarifies transfer rights in connection with a sale or transfer of the business or assets.
• It refers to subprocessors and international transfers using mechanisms such as the European Commission’s Standard Contractual Clauses.

5. Customer rights and contact details

• A specific privacy email address is added: privacy@sketch.com.
• The statement expressly mentions requests to update data, exercise privacy rights, and complain to a supervisory authority.