AI Policy

 
2025 PRETTYBIRD A.I. POLICY

This A.I. Policy shall govern all Employees and Freelancers that are hired by Prettybird and Prettybird’s affiliates.  Please direct any and all questions regarding this Policy to the legal department and the I.T. department.  Below are the guidelines and best practices for using A.I. Platforms in connection with Work created for Prettybird.

SECTION 1 – DEFINITIONS:
 1.1 – A.I. Platform(s): Any artificial intelligence platforms, models, programs, or software that may retain and/or use uploaded Data to train the artificial intelligence that is incorporated in such platforms, programs, or software (e.g. OpenAI, Midjourney, Google Gemini, Claude, etc.), and can be accessed by the general public. This also includes Generative A.I. platforms and models. For the avoidance of doubt, this would include any Prettybird created artificial intelligence platforms, models, programs, or software which is not accessible by the general public.
 1.2 – A.I. Privacy Policy: Guidelines on protecting Confidential Information and Data when using Prettybird approved Public A.I. Platforms. 
 1.3 – Confidential Information: Any information that is labeled as confidential or is information that an ordinary person would reasonably perceive as confidential.  This includes, but is not limited to, Prettybird Data, Data provided by a Third Party, Work, Proprietary Information, Trade Secrets, and PII.
 1.4 – Data: Treatments, scripts, concepts, ideas, financial information, Confidential Information, bids, budgets, art, photos, video clips, and any other form of documentation, image, or video.
 1.5 – Employee(s): Full-time, part-time, and temporary employees that are hired by, or on behalf of, Prettybird.
 1.6 – Freelancer(s): Independent contractors that are hired by, or on behalf of, Prettybird.
 1.7 – Generative A.I.: A.I. Platforms that have the capability of generating images or videos based off of prompts or uploaded Data.
 1.8 – NDA: Non-Disclosure Agreement.
 1.9 – Policy: Prettybird A.I. policy.
 1.10 – Personally Identifiable Information (“PII”): Personally Identifiable Information is information that can be used either independently or in conjunction with other information that would identify, or would be directly linked to, a specific individual.
 1.11 – Prettybird: Prettybird LLC with an address of 9905 Jefferson Blvd., Culver City, California 90232.
 1.12 – Prettybird Approved A.I. Platform List: A list of Prettybird approved Public A.I. Platforms and tools that are maintained by the I.T. department which lists all specifically approved Public A.I. Platforms that may be used by Employees and Freelancers. This list is subject to ongoing updates, change, and should be referred to before using any Public A.I. Platform in connection with employment under Prettybird. 
 1.13 – Prettybird A.I. Platform: An A.I. Platform owned, created, and controlled by Prettybird that cannot be accessed by the general public.
 1.14 – Prettybird Selected Representatives: Prettybird Employees who have the authority to bind Prettybird and Employees to the terms of an agreement, such as an NDA.
 1.15 – Project A.I. Database: A database that is maintained by the internal development team which outlines and catalogs in detail the specific Prettybird approved A.I. Platforms that are used in connection with each individual project (regardless of whether or not the project is fully awarded to Prettybird).  This is to ensure that Prettybird maintains a record of which A.I. Platforms are used and the manner in which these A.I. Platforms are used.
 1.16 – Proprietary Information: Any information that is owned by Prettybird or a Third Party, including, but not limited to, know-how, business operations, research and development methods, company databases, and Trade Secrets.
 1.17 – Public A.I. Platform(s): Any A.I. Platform not owned, created, or controlled by Prettybird that can be accessed by the general public, is connected to the internet, and/or uses off site servers to run the A.I. Platform (e.g. OpenAI, Midjourney, Google Gemini, Claude, etc.)
 1.18 – Third Party: Any company, agency, artist, person, music label, or management company that is, or has an interest in, hiring Prettybird to perform production services.
 1.19 – Trade Secrets: Any confidential and proprietary information that is owned by Prettybird or an affiliated Third Party that is confidential, unknown, and intentionally not readily available to the general public or to competitors, which has the potential to provide an advantage to competitors. 
 1.20 – Work: All of Employee and/or Freelancer’s services and the results and proceeds of such services in connection with the services that are rendered for Prettybird. 

SECTION 2 – USE OF A.I. IN THE WORKPLACE
 2.1 – Unacceptable Uses of A.I. Platforms in the Workplace:
 2.1.1 Employees and Freelancers shall not upload Prettybird’s nor a Third Party’s confidential materials and/or documents to any Public A.I. Platform. Employees and Freelancers shall not provide prompts that include Prettybird or Third Party Confidential Information to a Public A.I. Platform. This includes, but is not limited to uploading creative briefs, prompts, concept art, storyboards, descriptions, decks, or treatments that are provided by a Third Party to be used in the Work or to ideate on the Work. 
 2.1.2 Employees and Freelancers may not use any A.I. Platforms to create a summary of any Confidential Information.
 2.1.3 Employees and Freelancers may not use any Public A.I. Platforms that are not on the Prettybird Approved A.I. Platform List without express permission from their direct supervisor, unless Employees and/or Freelancers are directed to use a specific Public A.I. Platform by their direct supervisor or applicable Third Party.
 2.2 – Acceptable Uses of A.I. Platforms in the Workplace:
 2.2.1 Although the following are not considered Public A.I. Platforms, but rather artificial intelligence that has been embedded into commonly used programs, software, and websites, Employees and Freelancers are permitted to the limited use of auto fill suggestions in search engines, emails, texts, and other forms of communication, as well as spelling and grammar checking programs that are specifically associated with the underlying program or application, and such information is fully derived from publicly available information.
 2.2.2 Employees and Freelancers may use a Prettybird A.I. Platform where there is no public access to the Confidential Information that is entered into it. In no instance should Confidential Information be used on a Public A.I. Platform unless it is explicitly requested or approved by the applicable Third Party or Employee’s direct supervisor. 
 2.2.2.1 For the avoidance of doubt, Employees do not need prior approval from their direct supervisor, or the applicable Third Party, to use a Prettybird A.I. Platform or a Public A.I. Platform on the Prettybird Approved A.I. Platform List, unless it is known that the applicable Third Party who is engaging Prettybird does not allow the use of A.I. Platforms at any stage of a project .  
 2.2.2.2 Freelancers must receive approval from their direct supervisor, or the applicable Third Party, before using any A.I. Platform, including a Prettybird A.I. Platform or platforms on the Prettybird Approved A.I. Platform List.
 2.2.3 Employees and Freelancers may use Public A.I. Platforms in their Work as long as the information, documents, art, images, and videos are either made widely available to the public or in the public domain.
 2.2.4  Employees and Freelancers may use Public A.I. Platforms in their Work only if the prompts, files, and/or Data that are provided to the Public A.I. Platform does not include any Confidential Information or would contain any specific references to the Confidential information. For the avoidance of doubt, these prompts must be as generic as reasonably possible for the performance of the Work.
 2.2.5 If it is necessary to refer to a specific project by its name or title, such name or title must be anonymized with a codename when submitting such information into Public A.I. Platforms, and the code name must have no connections to the project itself.
 2.2.6 Employees and Freelancers may use Public A.I. Platforms in connection with the Confidential Information in the sole instance that the Third Party, whose Confidential Information would be shared with the Public A.I. Platforms, provides explicit, prior written approval prior to entering the Confidential Information into the Public A.I. Platforms.
 2.2.7 For the avoidance of doubt, in no situation is it acceptable to use either a Prettybird A.I. Platform or Public A.I. Platforms (even if on the Prettybird Approved A.I. Platform List) in connection with any Work created for a Third Party that is known to not allow the use of any A.I. Platforms at any stage of a project.
 2.3 – A.I. Platform Use Outside of the Workplace:
 2.3.1  This Policy in no way restricts Employees or Freelancers from using Public A.I. Platforms in their personal projects outside of their employment with Prettybird and on their own personal devices.
 2.3.2 For the avoidance of doubt, Employees and Freelancers may not use a Prettybird A.I. Platform for their personal projects outside of their employment with Prettybird. Employees and Freelancers are only permitted to use a Prettybird A.I. Platform for the sole purpose of creating the Work. 
 2.4 – Consequences of Violating This Policy:
 2.4.1  If an Employee or Freelancer violates this Policy, even if unintentional, they shall be subject to immediate disciplinary action, and such act shall be considered a breach of Employee and Freelancer’s nondisclosure obligations under their employment with Prettybird.
 2.4.2  If an Employee or Freelancer violates this Policy in any manner, such violation may open up themselves, as well as Prettybird, to legal action from the Third Party whose Confidential Information has been breached. 

SECTION 3 – A.I. PRIVACY POLICY AND BEST PRACTICES
 3.1 – Privacy Obligations For Confidential Information:
 3.1.1  All Employees and Freelancers are bound by the NDA that is signed when they are hired by Prettybird. Employees and Freelancers may also be bound by Third Party NDAs on a job-by-job basis as part of their employment under Prettybird. 
 3.1.2  Prettybird Selected Representatives may sign NDAs on behalf of themselves as well as Prettybird and Prettybird’s Employees. In certain situations, a Third Party may require that each individual sign a separate NDA as well; the highest confidentiality standard shall govern in such instance.
 3.1.3  Any Confidential Information that is disclosed to Employees and Freelancers is on a “need-to-know” basis.  This Confidential Information should not be shared with anyone who does not need access to such information in order to adequately perform the Work.
 3.1.4  If an Employee or Freelancer must share Confidential Information, it must be done securely. If possible and practicable, add password protection to the Confidential Information that is to be shared in tandem with Prettybird guidelines.
 3.1.5  If a Third Party requests or approves the use of a Public A.I. Platform in connection with their Confidential Information, Employees and Freelancers must strictly abide by the parameters that are approved by the Third Party, at all times subject to their direct supervisors’ approval.
 3.2 – Privacy Best Practices:
 3.2.1  Always opt-out of allowing the Public A.I. Platform to use Data that is uploaded for training and general online web usage. 
 3.2.2  When using Public A.I. Platforms on projects always opt-in for the maximum security options available and anonymize any references to products and/or projects with codewords.  
 3.2.3  When a project is completed, save the file in some manner (ex. a PDF) as a record and then delete the Public A.I. Platform’s history, if possible.
 3.2.4  If a log-in is required to use the Public A.I. Platform, do not use a personal log-in. Always use Prettybird’s verified log-in credentials if available. Prettybird’s credentials should only be kept in Bitwarden or per the Prettybird approved guidelines. This allows for Prettybird to monitor the Data that is input into the Public A.I. Platform and such violations of any Data that is input into the Public A.I. Platform shall be monitored and reported on.
 3.2.5  Only use Public A.I. Platforms that are on the Prettybird Approved A.I. Platform List.
 3.2.6  For each project (whether or not awarded), the Project A.I. Database, as well as the appropriate field in the corresponding job card in the Prettybird job tracker, should always be updated to ensure there is an accurate representation of all A.I. Platforms that are used in connection with the Work.
 3.3 – Information Collected by Public A.I. Platforms:
 3.3.1  Public A.I. Platforms automatically collect IP addresses, general locations, type of browser used, settings, user agent, date and time of access, interaction data (i.e. clicks, scrolls, features used), device type, operating system, and type of connection.
 3.3.2  Public A.I. Platforms may track preferences to personalize content through the use of cookies. This information may be shared with other parties such as Google, Meta, LinkedIn, Reddit, etc.
 3.3.3  From user input, Public A.I. Platforms can collect names, contact info, account credentials, date of birth, payment info, transaction history, input data, prompts, files, and even communication outside of the platform.

SECTION 4 – PUBLIC A.I. PLATFORM SPECIFIC PRIVACY GUIDELINES
 4.1 – ChatGPT/OpenAI
 4.1.1 Turn off data sharing with OpenAI. this prevents chats from being used to train the A.I. Platform. 
 4.1.1.1 If using OpenAI while being signed in: click the “Profile” icon, then click “Settings,” then “Data Controls,” and turn off “Improve the model for everyone” or the relevant equivalent to prevent training OpenAI’s platform.
 4.1.1.2 If using OpenAI on a mobile device, open the sidebar, click the “Profile” icon then click “Data Controls” and then turn off “Improve the model for everyone” or the relevant equivalent to prevent training OpenAI’s platform 
 4.1.1.3 If using OpenAI while being signed out: Click the “? (Help)” icon, then click “Settings” and turn off  “Improve the model for everyone” or the relevant equivalent to prevent training OpenAI’s platform
 4.1.2 In every instance where use of a Public AI Platform is approved in connection with the Work, delete the relevant chat history. Chats are saved by default unless they are manually deleted, meaning these chats must be deleted every time.
 4.1.2.1 To delete chat history click “Chat History,” then “Select a Conversation,” then delete the chat.
 4.1.2.2 Deletion removes the chat from the account immediately and permanently from OpenAI systems within 30 days.
 4.1.3 Disable or manage cookies on Chrome by clicking the three dots (top right), then click “Settings,” then click “Privacy and Security,” then click “Cookies and Other Site Data” and choose “Block third-party cookies.”
 4.1.4 To export data and download all account and chat information while logged in, visit ChatGPT Account Settings, then click “Data Controls” and then export the data. This will be sent by email with the exported data file attached.
 4.2 – Google Gemini:
 4.2.1 Unique to Gemini: It allows for human reviewers to read, annotate, and process chats; and data is saved on OpenAI for three (3) years, even if activity is deleted.
 4.2.2 If personal information is not correct, or is not being used in an approved way, a legal issue report can be submitted to Google.
 4.2.3 General Google privacy settings may be changed by accessing “Google Privacy Checkup.” 
 4.2.4 At all times disable data review and storage in Gemini by accessing Gemini’s “Activity Settings” and turn off “Gemini Apps Activity.”
 4.2.5 To export or delete data across Google in general, access “Google Privacy Tools.” Automated deletion of data can be set with these tools.
 4.3 – Midjourney:
 4.3.1 Unique to Midjourney: Midjourney does not guarantee any level of security, and there is no specified time frame in which they would delete past data.
 4.3.2 Control image visibility by using “Stealth Mode.” Please note that this is only available to “Pro” and higher tier users.
 4.3.2.1 Stealth Mode allows for images to be kept private or to unpublish after the images are generated but does not remove the image from Midjourney’s internal systems.
 4.3.3 Manage or withdraw cookies consent by accessing the “Privacy Settings.” Please note, that cookies cannot be fully disabled.
 4.3.4 Data can be accessed, corrected, or deleted by submitting a request through the “Midjourney Help Center” or the appliable equivalent. 
 4.4 – Claude (Anthropic):
 4.4.1 Customer data that is stored on Claude’s platform is associated with specific customers, but this information is separated from other customers data and the internet.
 4.4.2 Data entered into Claude is encrypted can only be accessed by those with the proper credentials. Credentials should only be given to Employees who are approved to use Claude by their direct supervisor.
 4.4.3 By default, Claude will not use any data entered into chats or sessions when using their commercial products. The only time data would be used to train an A.I. Platform is if the customer explicitly opts-in.
 4.4.4 Data is retained within the “Claude for Work” platform until it is deleted by the customer. After the Data has been deleted, it may stay in their system for up to thirty (30) days. If requested, they will issue a certificate of destruction.

Updated 2025-12-08