Invisible Technologies

Summary of Important Changes

1. Scope of the Privacy Policy Expanded or Clarified

• The policy now states that it applies to personal information processed by Invisible in its business, including its websites, mobile applications, and related online and offline offerings.
• The prior wording appears to have been limited to saying the policy “applies to” certain information; the new language more expressly covers information processed across the company’s business activities.
• The policy now expressly states that Invisible may collect, use, and share personal information and assist individuals in exercising applicable privacy rights.

Risk: The broader scope may bring more types of customer, agent, website, and project-related information within the policy’s stated coverage.

2. Data Sharing Language Is More Explicit

The revised policy describes sharing personal information with a broad range of recipients, including:

• Clients and customers;
• Service providers;
• Business partners and affiliates;
• Advertising partners;
• Providers of APIs and software development kits;
• Parties involved in mergers, acquisitions, financing, or asset transfers; and
• Governmental or law-enforcement authorities in specified circumstances.

Risk: Customer data may be disclosed to numerous third parties and corporate affiliates. The policy does not, in the provided changes, materially narrow the purposes, categories of data, or contractual restrictions applicable to those recipients.

3. AI Model Training: No Clear New Customer-Data Training Right Identified

• The diff does not clearly add language expressly authorizing Invisible to use customer data, project data, work product, prompts, outputs, or personal information to train, fine-tune, or improve AI models.
• The phrase “value for AI” appears as part of a substantial website-content replacement, but it does not itself create a clear data-use permission.
• The policy continues to refer to:
• “Agent training”;
• Research and development;
• Statistical analysis;
• De-identified and aggregated information; and
• Other legally permissible purposes.

Risk: These broad existing purposes could potentially support certain model-development or analytics activities, particularly using de-identified or aggregated information. However, the policy does not define whether customer-provided content or project work product is de-identified, whether it is retained for model training, or whether customers can opt out.

Recommended clarification: Customers should seek an express contractual statement addressing whether their data may be used for AI training, whether data is excluded by default, what anonymization standards apply, retention periods, human review, and opt-out/deletion rights.

4. Cookies, Analytics, and Advertising Controls

• The revised content more clearly describes operational, performance, functionality, advertising, targeting, and cross-device tracking technologies.
• Cookie-consent language now includes “Reject All,” “Accept All,” and preference-management options.
• Google Analytics and other third-party analytics providers remain contemplated.

Risk: Website usage data may be used for analytics, personalization, targeted advertising, and cross-device tracking, subject to jurisdiction-specific rights and consent mechanisms.

5. Administrative and Presentation Changes

• The policy includes a privacy email address and updated “last reviewed” information.
• Much of the diff consists of navigation, footer, formatting, and website-content changes rather than substantive legal amendments.
• No clear changes were identified to liability disclaimers, security commitments, retention periods, international-transfer mechanisms, or children’s-data provisions.

Summary

The diff only states that approximately 184 words were removed; it does not identify the deleted language or provide a clean version of the revised document.

Key Legal Implications

• Unassessable substantive changes: Without the actual deleted text, it is not possible to determine whether the changes affect liability, confidentiality, privacy, intellectual property, security, termination, or other contractual obligations.
• Potential loss of customer protections: Deleting language could remove or narrow customer rights, warranties, service commitments, audit rights, indemnities, data-security obligations, or restrictions on the provider’s use of information.
• AI-model training: The available diff does not show whether the agreement now permits, restricts, or remains silent on using customer data to train, fine-tune, evaluate, or improve artificial-intelligence or machine-learning models. This issue cannot be analyzed from the word-count description alone.
• Interpretation risk: If the deleted provisions addressed customer data, the deletion may create ambiguity about whether data can be used for purposes beyond providing the contracted services.

Information Needed

To perform a meaningful legal review, provide either:

1. The actual deleted 184 words, marked with deletion brackets; or

2. The prior and revised versions of the relevant provision or document.

Particular attention should be given to language concerning:

• Use of customer content or personal data;
• AI or machine-learning model training and improvement;
• Data retention and deletion;
• De-identification or aggregation;
• Disclosure to affiliates or third-party vendors; and
• Ownership and license rights.

Structured Summary of Important Changes

1. Overall nature of the changes

The diff is largely a website/content restructure, including:

• Reorganized navigation, industry labels, offerings, and footer content.
• Formatting changes removing spaces and paragraph breaks.
• Updated privacy-policy table of contents and cookie-consent interface.
• New or revised contact information, including privacy@invisible.email.
• Changes to the policy’s introductory “Scope” language.

Many apparent text changes are caused by website navigation or scraping artifacts rather than substantive contractual amendments.

2. Scope of the Privacy Policy

The revised language more clearly states that the Privacy Policy:

• Applies to personal information processed by Invisible in its business.
• Covers websites, mobile applications, and related online and offline offerings, collectively defined as the “Services.”
• Explains how Invisible collects, uses, and shares personal information and assists individuals in exercising privacy rights.

Risk: The scope remains broad and applies across multiple types of users, including Clients, Agents, and other individuals. Customers should confirm whether separate customer contracts, data-processing agreements, or project-specific terms impose stricter limits.

3. Customer data use and AI-model training

No express new model-training authorization identified

The diff does not clearly add language stating that customer data, customer content, project data, or work product may be used to train Invisible’s or third parties’ AI models.

The revised website includes marketing language referring to:

• “value for AI”
• Frontier model-building research
• Intelligent enterprise deployment

However, these appear to be company-description or navigation content, not an operative data-use provision.

Existing provisions that may still create indirect AI-related risk

The policy continues to describe broad uses of personal information, including:

• Research and development.
• Statistical analysis and analytics.
• Creation and use of de-identified or aggregated information for research, internal analysis, analytics, and other legally permissible purposes.
• Agent training “or other purposes.”
• Automated decision-making and profiling.
• Sharing with service providers, affiliates, business partners, APIs, and SDK providers.

Risk: Although not expressly authorizing model training, these broad provisions could potentially be interpreted to permit use of de-identified, aggregated, or otherwise permitted data in developing or improving AI systems. The policy does not clearly define:

• Whether customer-provided content is included.
• Whether data may be used to train general-purpose models.
• Whether prompts, outputs, annotations, or work product may be retained.
• Whether third-party AI providers may use the data for their own training.
• Whether customers can opt out or require deletion.

Customers handling confidential, regulated, or sensitive data should seek an express contractual restriction on AI training and confirm applicable data-processing terms.

4. Cookies and tracking

The revised cookie-consent content adds or clarifies categories for:

• Essential technologies.
• Analytics.
• Marketing and targeted advertising.
• Personalization.
• Cross-device tracking.

It also adds clearer “Reject All,” “Accept All,” and preference-management language.

Risk: The policy continues to permit broad use of analytics, advertising technologies, third-party providers, and cross-device tracking, subject to available opt-outs.

5. Other notable issues

• International data transfers remain broadly permitted.
• Data retention language remains general and does not establish precise customer-data deletion periods.
• Security language includes a broad limitation of liability for unauthorized disclosure to the extent permitted by law.
• The policy may be changed over time, with continued use treated as acceptance.
• The revised footer appears to include a blank or incomplete “Last reviewed” field, which may create version-control uncertainty.

Structured Summary of Important Changes

1. Privacy policy scope expanded and clarified

• The policy now applies to personal information processed by Invisible in its broader business, including:
• Websites;
• Mobile applications; and
• Related online and offline offerings, collectively defined as the “Services.”
• The policy now expressly describes Invisible as outsourcing business processes to human agents.
• The categories of information collected are stated to depend on how the individual interacts with the Services and applicable law.

Risk: The broader scope may bring additional products, channels, and offline activities within the policy without requiring a separate notice. Customers should confirm that any contractual data-processing restrictions apply across all such Services.

2. More detailed data collection disclosures

The policy continues to identify extensive categories of personal information, including:

• Account, contact, payment, tax, government identification, and background information;
• Project information and work product;
• Communications, support interactions, surveys, and event information;
• Agent work information and training-related information;
• Cookies, analytics, advertising, cross-device tracking, and information from third parties.

Risk: The breadth of the categories—particularly project information, work product, communications, and government identification—may encompass customer-provided data and confidential business information.

3. Broader stated purposes for use

The policy permits use for:

• Providing and administering Services;
• Security, fraud detection, legal compliance, and enforcing agreements;
• Marketing, research and development, analytics, and strategic partnerships;
• Automated decision-making and profiling;
• Creating de-identified and aggregated information for research, internal analysis, analytics, and any other legally permissible purposes.

Risk: The broad “research and development” and “any other legally permissible purposes” language could support secondary uses beyond the immediate customer engagement, subject to applicable law and any separate customer contract.

4. AI-model training implications

• The diff does not expressly add a clause stating that customer data, customer content, or project work product may be used to train generative AI or machine-learning models.
• It does retain or present language concerning “Agent training,” but this appears directed to training human agents or collecting information for client training purposes—not expressly training AI models.
• The policy refers to Invisible powering AI and to research and development, automated decision-making, and de-identified/aggregated data use. These provisions could potentially support AI development or model improvement, particularly for data treated as de-identified or aggregated.
• There is no clear customer-data opt-out, retention limit, model-use restriction, or express promise that customer content will not be used for AI training.

Key risk: The absence of an express AI-training authorization or prohibition creates ambiguity. Customers should seek contractual confirmation regarding whether their data, prompts, outputs, work product, or metadata may be used to train, fine-tune, evaluate, or improve AI models.

5. Expanded disclosure and tracking language

The policy continues to permit sharing with clients, service providers, affiliates, business partners, advertising partners, APIs/SDK providers, and transaction counterparties. Cookie language also covers targeted advertising, analytics, personalization, and cross-device tracking.

Risk: Customer or employee data may be disclosed to a wider vendor ecosystem, including international recipients and technology providers, subject to applicable law.

6. Administrative updates

• The policy is dated March 25, 2026.
• An email privacy contact was added.
• Formatting, navigation, cookie-consent language, and website content were substantially reorganized.

These changes appear primarily administrative, but the expanded website and Services scope should be reviewed against customer agreements.

Summary

The diff only states that approximately 184 words were removed. The actual deleted language is not provided.

Key Legal Implications

• Scope of review is limited: It is not possible to determine which obligations, rights, limitations, or protections were removed.
• Potentially material deletions: The removed text could have affected important provisions such as:
• Customer data ownership and permitted uses
• Confidentiality and security obligations
• Use of customer data to train, fine-tune, or improve AI models
• Restrictions on sharing data with service providers or third parties
• Data retention and deletion
• Intellectual-property rights and ownership of outputs
• Warranties, indemnities, liability limitations, or termination rights
• Compliance with privacy and data-protection laws

AI Training and Data Use

No conclusion can be reached about changes to AI-model training because the deleted language is not included. In particular, the available diff does not show whether the contract:

• Newly permits or removes permission to use customer data for training or improving AI models;
• Limits training use to de-identified, aggregated, or anonymized data;
• Requires customer consent or provides an opt-out;
• Restricts human review or use of data for model evaluation;
• Addresses whether prompts, inputs, outputs, or usage metadata may be retained or used for training; or
• Requires deletion of customer data from training datasets or model-related systems.

Risk Assessment

The deletion may create uncertain or increased risk if it removed restrictions on data use, confidentiality protections, or commitments not to train models on customer data. Conversely, it could reduce unnecessary or conflicting language if the deleted provisions were duplicative or replaced elsewhere.

A reliable legal analysis requires the text of the approximately 184 deleted words and, ideally, the surrounding unchanged provisions.

Structured Summary of Important Changes

1. Privacy Policy Scope Expanded and Reframed

• The policy is now presented as applying to personal information processed in Invisible’s business, including:
• Websites;
• Mobile applications; and
• Related online and offline offerings, collectively defined as the “Services.”
• The prior wording focused more narrowly on information collected through the Services.
• The policy now states that information categories depend on how individuals interact with the Services and the requirements of applicable law.

Risk: The expanded scope may bring additional business activities and data-processing contexts within the policy without clearly identifying the applicable purposes, legal bases, retention periods, or customer-specific controls for each activity.

2. AI, Model Development, and Training-Related Language

The diff does not add a clear provision expressly stating that customer-provided data, Project data, work product, prompts, outputs, or other Client content may be used to train general-purpose AI models.

However, the policy continues to include broad purposes that could potentially encompass AI development or model improvement, including:

• Research and development;
• Statistical analysis;
• Internal analysis and analytics;
• Creation and use of de-identified or aggregated information for “any other legally permissible purposes”;
• Agent training by Clients or “for other purposes”; and
• Automated decision-making and profiling.

The updated website branding also references “AI,” “frontier model-building research,” and “intelligent enterprise deployment,” but these additions do not themselves create a precise contractual authorization or limitation concerning customer data.

Key risks and ambiguities:

• No clear distinction between:
• Customer data used to perform a contracted project;
• Data used to improve Invisible’s services;
• Data used to train internal or third-party AI models; and
• De-identified or aggregated data used for research.
• No express opt-out, consent mechanism, retention limit, model-unlearning obligation, or prohibition on using confidential customer content for model training.
• “De-identified,” “aggregated,” and “legally permissible purposes” are broad and undefined in the excerpt.
• If this policy is intended to govern customer data, it may be insufficient by itself to establish appropriate data-processing rights, particularly for confidential, regulated, or personal data.

3. Data Collection and Disclosure Structure Clarified

The policy organizes data collection by category, including:

• Account and Client information;
• Agent information;
• Project information and work product;
• Payment information;
• Communications, surveys, support interactions, and event information;
• Automatically collected data, cookies, analytics, and cross-device tracking; and
• Information obtained from other sources.

Disclosure categories are also stated more explicitly, including Clients, service providers, business partners, affiliates, advertising partners, APIs/SDKs, and parties involved in corporate transactions.

Risk: The broad third-party disclosure language may permit extensive sharing unless customer contracts impose stricter confidentiality, security, subprocessors, and use restrictions.

4. Cookies, Advertising, and Tracking

The revision adds or clarifies:

• Analytics, advertising, targeting, personalization, and cross-device tracking;
• Cookie preference controls, including “Reject All” and “Accept All”; and
• Consent language stating that acceptance permits cookie storage to enhance services and support marketing.

Risk: Tracking and advertising practices may create additional consent, opt-out, “sale/share,” profiling, and international privacy compliance obligations.

5. Rights, Retention, and Security

The policy retains general rights, international-transfer, retention, security, supervisory-authority, and children’s-information sections. It adds a contact email and updates the policy presentation/date to March 25, 2026.

Risk: The excerpt still does not provide specific retention periods, detailed security commitments, or customer-specific deletion and return obligations.

No

Summary of Important Changes

1. Privacy policy presentation and scope

• The policy has been reformatted and expanded with clearer headings, navigation, and spacing.
• The scope is now stated more explicitly: the policy applies to personal information processed by Invisible in its business, including its websites, mobile applications, and related online/offline offerings.
• The introductory language now expressly says the policy explains how Invisible may collect, use, and share personal information and helps individuals exercise applicable privacy rights.
• The contact section now adds an email address and revised “last updated/last reviewed” fields.

Risk: The revised scope is broad and may apply across multiple products, services, websites, and business activities. Customers should not assume that project data is governed only by the specific service agreement or statement of work.

2. AI-model training and use of customer data

• The diff does not clearly add an express provision authorizing Invisible to use customer data, project data, work product, prompts, outputs, or personal information to train general-purpose AI or machine-learning models.
• The policy continues to describe:
• “Agent training” and collection of information for training agents or other purposes;
• research and development, analytics, and marketing research;
• creation and use of de-identified and aggregated information for research, internal analysis, analytics, and “any other legally permissible purposes.”
• These provisions appear substantially unchanged in substance, although the revised formatting makes them more prominent and easier to identify.

Key risk: The broad research, development, analytics, and de-identification language could potentially be interpreted to permit use of customer-related information for improving AI systems, developing models, or other internal technologies—particularly if the information is treated as de-identified or aggregated. However, the policy does not define:

• whether customer content or work product is included;
• whether data may be used to train foundation or third-party models;
• whether models may be trained for other customers;
• whether customers can opt out;
• how re-identification risk is managed; or
• how long training data or model-derived information is retained.

Customers requiring “no training” treatment should obtain an express contractual restriction, including a prohibition on using customer data, prompts, outputs, metadata, and derived information for model training or product improvement, except with written consent.

3. Broader sharing and tracking disclosures

• The policy continues to permit sharing with clients, service providers, affiliates, business partners, advertising partners, API/SDK providers, and parties involved in corporate transactions.
• Cookie, analytics, advertising, cross-device tracking, and personalization disclosures are presented in greater detail through the consent interface.

Risk: Customer or user information may be exposed to additional vendors or tracking technologies, subject to applicable consent requirements and contractual safeguards.

4. Other legal implications

• Automated decision-making and profiling remain disclosed without meaningful additional detail.
• International transfers, retention, security limitations, privacy rights, and policy-change provisions remain broadly permissive.
• The diff contains substantial website navigation and cookie-banner changes that appear non-substantive legally, but the rendered policy should be reviewed to ensure no accidental omissions or broken text.

Summary of Important Changes

1. Privacy Policy scope and description

• The opening description changes from Invisible delivering digital work “to value for AI” to delivering digital work “to human agents.” This may narrow or alter the description of the company’s business and could create ambiguity about whether AI-related processing is covered.
• The Privacy Policy is more clearly stated to cover how Invisible “collect[s], use[s] and share[s]” personal information.
• The scope is expanded or clarified to cover personal information processed in Invisible’s business, including its websites, mobile applications, and related online and offline offerings.
• The policy is dated March 25, 2026. The updated text appears to remove or omit the prior “Last updated” / “Last reviewed” presentation and the privacy email address. This may make it harder for individuals or customers to identify the effective date or submit privacy inquiries.

2. Uses of data

The revised policy continues to authorize broad uses of personal information, including:

• Providing services, administering accounts, processing payments, and customer support;
• Legitimate-interest activities such as direct marketing, research and development, and marketing research;
• Fraud detection, statistical analysis, internal analysis, analytics, and other legally permissible purposes;
• Tailored marketing, advertising, profiling, and automated decision-making;
• Creation and use of de-identified and aggregated information for research, analytics, internal analysis, and other legally permissible purposes.

These provisions remain broad and may permit secondary uses beyond delivering the customer-requested services. The policy does not appear to impose a specific deletion period or clearly limit derived or aggregated datasets.

3. AI-model training

• The diff does not add an express provision stating that customer data, Project data, work product, prompts, outputs, or personal information may be used to train, fine-tune, evaluate, or improve AI models.
• The policy does retain language concerning “Agent training” and the use of Agent work information for fraud detection, statistical analysis, or other purposes. This concerns training human agents or analyzing their work, but it is not clearly an authorization to train AI models.
• The broad references to research and development, analytics, automated decision-making, and de-identified or aggregated information could nevertheless support AI-related development or model improvement, particularly where data is claimed to be de-identified.
• Customers should seek an explicit contractual commitment addressing whether their data and work product may be used for AI training, whether identifiable or confidential data is excluded, whether opt-out rights exist, and whether any trained models or derivatives may be retained.

4. Other notable risk areas

• The policy continues to permit sharing with numerous categories of third parties, including service providers, affiliates, business partners, advertising partners, APIs, and SDK providers.
• Cross-device tracking, targeted advertising, analytics, and automated profiling remain contemplated.
• International transfers remain permitted.
• The revised presentation includes extensive cookie-consent and marketing language, but the substantive consent mechanism and provider disclosures should be verified for completeness and regulatory compliance.