Artificial Intelligence Addendum – KAGR LLC
Last Updated: 10/1/2026
This Artificial Intelligence Addendum (“AI Addendum”), forms part of the Master Client Agreement (as applicable, the “Agreement”) between KAGR LLC (“Provider” or “KAGR”) and the entity that has engaged Provider to provide the Services, including the Platform (“Client”). Each of Provider and Client is referred to in this AI Addendum individually as a “Party”, collectively the “Parties”. By entering into the Agreement, the Parties are deemed to accept, acknowledge and agree to all terms and conditions this AI Addendum. Capitalized terms used in this AI Addendum have the meaning ascribed to them in this AI Addendum or, in the event no definition is provided hereunder, in the Agreement. In the event of a conflict between the terms of this AI Addendum and a relevant SOW, the terms of the relevant SOW shall control.
1.Authorization.
a. Default. KAGR will not use large language models and other forms of generative or agentic AI in connection with the Processing of Client Content under the Agreement without Company’s prior written consent as set forth in this Addendum or a relevant SOW or unless it is specifically approved in writing in advance.
b. Authorization. Client hereby authorizes KAGR to use the AI Systems solely to carry out AI operations in connection with the provision of any Services or deliverables contemplated by, and/or for the authorized use-cases set forth in, the Agreement or a relevant SOW between KAGR and Client (the “Authorized Use-Cases”). The duration of the authorization for the Authorized Use-Cases shall be for the term of, or as otherwise expressly or implicitly permitted by, the Agreement or relevant SOW.
c. Use Restrictions. Client shall not, and shall not allow any Authorized User or third party who gains access to the Platform with Client’s authorization or due to Client’s failure to use reasonable security precautions to, use the AI Systems or the Platform for the following: (i) automated decision-making processes with legal or similarly significant effects, unless Client ensures that the final decision(s) is made by a human being and takes other factors beyond the Services’ recommendation into account; (ii) generating or providing individualized advice, including without limitation medical advice, treatment or diagnosis, that in the ordinary course of business would be provided by a licensed professional; (iii) explicitly predicting or categorizing based on an individual’s protected characteristic; provided, that this restriction does not limit or prohibit use cases designed specifically to identify and reduce bias in the AI Systems; (iv) social scoring or crime prediction based on social behavior or personality characteristics where such scoring leads to detrimental or unfavorable treatment unrelated to the original context of the Client Data, or unjustified or disproportionate treatment relative to the assessed behavior; (v) submitting images, videos or audio recordings of individuals without consent for the purposes of creating, analyzing or categorizing based on biometric identifiers; (vi) engaging in deceptive, manipulative or exploitative activities; or (vii) creating, sending, uploading, displaying, storing, processing or transmitting sexually explicit materials, sexual chatbots or materials that may be harmful to minors or engaging in erotic chats. If Client accesses and/or uses an AI System or AI from third-parties in connection with the Platform, Client is responsible for its, and its Authorized Users’, compliance with all terms and conditions and policies provided by the third-party AI System or AI provider with respect to the access and use of such third-party AI System or AI.
2.Ownership; Proprietary Rights.
a. Client Content. Client Content is solely owned by Client, including all rights, title and interest in and to Client Content, including any intellectual property, goodwill and other proprietary rights embodied therein or thereby. KAGR hereby assigns to Client all worldwide right, title, and interest that KAGR may have or acquire in or to Client Content, including any intellectual property, goodwill and other proprietary rights embodied therein or thereby. Client hereby grants to Provider a worldwide, non-exclusive, fully paid-up license to use the Client Content solely (i) in any manner reasonably necessary to provide the Platform to Client and/or perform the Services and/or (ii) in an anonymized form or in an aggregated format including with data provided by other third parties provided that it is not possible to identify Client or the identity of any individual contained in the Client Content. No right, license, permission, or ownership or other interest of any kind in or to the Client Content is or is intended to be given or transferred to or acquired by KAGR except as expressly stated in writing in the Agreement, and other than the rights expressly granted to KAGR in and to the Client Content, Client reserves all rights to the Client Content.
b. Proprietary Systems. All right, title, and interest in the intellectual property embodied in the AI Systems, Content provided by Provider through the AI Systems (other than Client Content), any customizations of the AI Systems, and any other adaptation, modification, derivation, addition or extension thereof (including but not limited to customized or adapted data via exports), will belong solely and exclusively to Provider or Provider’s licensors, and Client will have no rights in any of the above, except as expressly granted in the Agreement.
c. Notwithstanding anything to the contrary in the Agreement (including this AI Addendum), Client acknowledges that Provider shall have a right to use and disclose data relating to the operation, support and/or use of the AI Systems for its legitimate business purposes, such as product development and sales and marketing.
3.Representations and Warranties.
a. Without limitation of any other provision herein or in the Agreement, Client represents, warrants and covenants:
i. that Client has (and will have) collected and disclosed all Client Content in compliance with all Applicable Laws and provided any notice and obtained all consents and rights required by Applicable Law to enable Provider to lawfully use the AI Systems and otherwise Process Client Content as permitted by the Agreement and/or this AI Addendum;
ii. that Client has (and will continue to have) full right and authority to make the Client Content available to Provider under the Agreement and this AI Addendum;
iii. that Provider’s Processing of the Client Content in accordance with the Agreement, this AI Addendum, and/or Client’s instructions do not and will not infringe upon or violate any Applicable Laws, applicable privacy policies or the intellectual property rights or other proprietary rights of any other party; and
iv. that Client’s and its Custodians’ and Authorized Users’ use of the AI Systems and Platform will not involve unethical, harmful, deceptive, or illegal materials (including Training Data) or practices or violate Industry Standards or Industry Standard AI principles of accuracy, fairness (including to limit the risk of Discrimination), explainability to understand AI logic and decision-making, reliability, robustness, resilience, safety, and security (“AI Principles”).
b. KAGR represents and warrants:
i. that KAGR lawfully procured and/or possesses all relevant rights and consents for the Parties to use the AI System Content for the purposes contemplated hereunder in accordance with this AI Addendum and the Agreement;
ii. that the AI Systems were designed to, and will continue to, be used by KAGR in the connection with the Agreement in a responsible manner, including with considerations for protecting the privacy of individuals and relevant AI Principles;
iii. that, to KAGR’s knowledge, the AI Systems do not involve unethical, harmful, deceptive, or illegal materials (including training data) or practices;
iv. that, to KAGR’s knowledge, the AI Systems, any training data utilized by KAGR and Client’s use of the AI Systems have not infringed, misappropriated, or otherwise violated, and do not and will not infringe, misappropriate, or otherwise violate any third-party intellectual property rights, including but not limited to patents, copyrights, trademarks, trade secrets, or any other proprietary rights; and
v. to the extent required by Applicable Laws, to use reasonable development standards and processes to design, utilize, test and audit (as applicable) each AI System, for each applicable Authorized Use-Case, including to undertake the following relevant activities in good faith: (a) to select training data that is representative, not knowingly derived from deceptive or inaccurate data and inclusive of a statistically sufficient sample size and validate the same; (b) to regularly and routinely test and verify training data and Outputs, excluding Client Outputs, to confirm accuracy, completeness, validity, timeliness, the lack of inherent Discrimination, Memorization, Hallucinations, or Data Leakage; (c) to prevent the AI Systems from taking into consideration actual or inferred protected characteristics, including: race, ethnicity, sexual preference, religion, national origin, disability, gender, marital status or age of individuals, except, if applicable, for specific purposes that have been pre-approved by Client in writing; and (d) to prevent the AI Systems from developing otherwise offensive, harassing, discriminatory, harmful or inappropriate behaviors;
provided, however, that the representations and warranties set forth in this Section 3(b) do not apply, and KAGR disclaims any express or implied warranties, with respect to (1) modifications of or to the AI Systems, inclusive of the AI System Content, designed and/or implemented by or on behalf of Client; (2) Client’s use of AI Systems, inclusive of the AI System Content, that are no longer supported by KAGR (or its licensors) or in an unauthorized manner; (3) Client Content or Outputs; (4) third-party AI Systems or AI, or Client’s use thereof; or (5) infringement claims that are predicated on Client Content infringing on, misappropriating or otherwise violating third-party rights, including without limitation intellectual property rights.
4. Kill Switch. Except with respect to third-party AI Systems or other AI used by third parties, KAGR represents, warrants and covenants that each AI System or application of AI in connection with the Agreement, in each case in connection with the Agreement, has and will have (i) a technological mechanism to prevent the AI System (or a component thereof) or application of AI from executing or completing any or all of its functions; or (ii) another emergency or failsafe mechanism (including human intervention as a failsafe) to prevent the AI System (or a component thereof) or application of AI from executing or completing its functions (“Kill Switch”). KAGR agrees to take any commercially reasonable steps to enable the Kill Switch in response to an AI Issue if KAGR deems appropriate. KAGR shall use good faith efforts to minimize any materially negative impact to Client’s use of the Platform Services in the event of any use of the Kill Switch and to continue to perform the Services and provide the deliverables contemplated by the Agreement. In the event that KAGR uses the Kill Switch with respect to any AI System (or a component thereof) or application of AI utilized in connection with the provision of Services or deliverables under the Agreement, KAGR will provide immediate notice of such use to Client.
5. Risk Management
a. Compliance with Applicable Law. Each Party will provide the other Party with commercially reasonable assistance during the Term to enable the other Party to comply with Applicable Law in relation to its use of the AI Systems (or outputs thereof).
b. Change in Applicable Law. The Parties acknowledge and agree that if, after the date hereof, any Applicable Law applicable to a Party’s use of the AI System is modified, amended, adopted or otherwise becomes effective during the Term, to the extent the terms of this AI Addendum conflict with such new or modified Applicable Law that takes effect during the Term, the Parties shall, upon such Party’s request, negotiate in good faith to amend the terms of this AI Addendum to the extent necessary to comply with such Applicable Law.
c. Notification of AI Issue. Subject to any directions from law enforcement to the contrary, KAGR will notify Client promptly (and in no event more than 72 hours) after becoming aware of any actual or suspected AI Issue affecting Client Content, so that Client can fulfil any AI-related incident reporting obligations it may have under Applicable Law. KAGR will further take reasonable and appropriate measures and actions to minimize the impact of the AI Issue (to the extent such AI Issue was caused by KAGR’s acts or omissions) and shall keep Client informed of all material developments in connection with the AI Issue to the extent known by KAGR. KAGR will provide reasonable assistance to, and shall reasonably cooperate with all reasonable requests of, Client to investigate and (to the extent such AI Issue was caused by KAGR’s acts or omissions) to mitigate and/or address such AI Issue.
d. Human Oversight. Each Party will use commercially reasonable efforts to ensure that its use of an AI System in connection with the Agreement is subject to human oversight at a level that is reasonable in light of the expected use of such AI System.
6. Audit. KAGR shall maintain a reasonably complete audit trail of all Platform activities associated with the use of the AI Systems and AI in connection with the Agreement, which may include without limitation model inferences, errors and factors affecting outputs. Promptly upon Client’s request, KAGR will provide Client with more detailed information regarding KAGR’s AI risk management program. KAGR will, upon Client’s reasonable prior written request, provide an independent and nationally recognized auditor (which auditor shall not be a competitor of KAGR or compensated on a contingency basis) with access to KAGR’s systems and records that involve or are related to any Processing of Client Content by AI Systems so that an audit may be conducted. Client may not exercise such audit right more frequently than once per twelve (12) month period and Client will bear the full cost and expense of any such audit, provided that if such audit discloses a AI Issue, the auditor shall disclose all details of such AI Issue to KAGR, and, to the extent such AI Issue was caused by KAGR’s acts or omissions, KAGR will bear the reasonable costs and expenses of such audit and a further independent audit may be conducted within the then-current twelve (12) month period in Client’s discretion. No such audit shall be undertaken unless or until Client has requested, and KAGR has provided, information regarding KAGR’s AI risk program under this Section 6 and Client reasonably determines that such an audit remains necessary to demonstrate material compliance with the obligations set forth herein. Notwithstanding the foregoing, if KAGR provides Client with a current copy of its ISO/IEC 42001 (or substantially similar) certification, if any, performed by an independent nationally recognized auditor, such certification shall provide sufficient evidence that KAGR has a commercially reasonable AI risk management program in place, and Client shall not have the right to request a security audit of KAGR. In no event shall Client or any auditor have access to the information of any other client of KAGR and the disclosures made pursuant to this Section 6 (“AI Audit Information”) shall be held in confidence as KAGR’s Confidential Information and subject to any confidentiality obligations in the Agreement. Without limiting the generality of any provision in the Agreement, Client shall employ the same degree of care to safeguard AI Audit Information that it uses to protect its own confidential and proprietary information and in any event, not less than a reasonable degree of care under the circumstances, and Client shall be liable for any improper disclosure or use of AI Audit Information by Client, an auditor or Client’s agents. Further, Client shall ensure that any third-party auditor under this Section 6 is bound by obligations of confidentiality broad enough to encompass AI Audit Information that are at least as protective thereof as the Agreement and this Section 6.
7. Definitions. As used in this AI Addendum:
“AI” means any product, service, software, algorithm, system, model or tool that relies on or otherwise utilizes artificial intelligence, machine learning, large language models, or other similar or successor technologies, in whole or in part, to create, generate, modify or output data, Content, or information of any kind.
“AI Issue” means an event, occurrence or issue with an AI System or any other application of AI in connection with the Agreement that KAGR or Client knows or reasonably suspects (i) causes or will foreseeably cause the production of inaccurate, unfair, Discriminatory or unrepresentative Client Outputs, or Client Outputs that consider actual or inferred protected characteristics, (ii) will violate or will foreseeably violate material Industry Standards and/or AI Principles applicable to this AI Addendum, or (iii) results in, is reasonably likely to result in, or arises as a consequence of Hallucinations, Memorization, or Data Leakage.
“AI System” means any product, Service, system, deliverable, software, algorithm, tool, semantic layer or feature involving AI that KAGR licenses or makes available to Client under or in connection with the Agreement.
“Client Content” means all Client Data, Client Inputs, Client Outputs, Training Data, or Confidential Information of Client, in any form or format, including interim, Processed, compiled, summarized, or derivative versions thereof, that may exist in any system, tool, product, service, algorithm, tool, feature, database, or record that is either (i) provided by or on behalf of Client or its customers or partners to KAGR, or (ii) is obtained, developed, produced or Processed by KAGR or KAGR systems, in each of (i) and (ii) in connection with the relationship or arrangements established by the Agreement, but excluding any deliverables, data, Content, or information, that is expressly owned by KAGR pursuant to the Agreement. Any successors, equivalents, compilations or derivatives of the foregoing, whether now known or hereafter devised, and in any medium or format, are also Client Content. For example, copying or tracking of any portion of Client Content to create a separate set of data, Content, or information, constitutes a derivative and is within the definition of Client Content.
“Client Input” means Inputs used in connection with the Agreement or otherwise by or on behalf of Client.
“Client Output” means any Output generated in connection with the Agreement or otherwise by or on behalf of Client, or based on Client Inputs.
“Client Training Data” means any Training Data that is provided by or on behalf of Client, or otherwise procured and used in connection with the Agreement.
“Content” means content, algorithms, artifacts (tokens, weights, files), models, context, prompts, insights, works of authorship, facts, images, videos, data, audio files and other materials.
“Data Leakage” means the exposure of non-public Client Content, or Client Content not intended to be exposed, by AI, including exposure, via an Output, of an Input from another user, non-public data, or Client Content that had been used to train the applicable AI.
“Discrimination” means, with respect to any AI System or any other application of AI, that the AI System or any other application of AI produces systematically incorrect, incomplete or inaccurate predictions or other Outputs or incorrect or misleading attribution and with respect to data, that the data is inaccurate, incomplete, unfair or otherwise not representative of the relevant population of interest. Client will have the right to determine the existence of any Discrimination or likely Discrimination in its discretion.
“Hallucinate” (including, with correlative meaning, “Hallucination”) means the creation of an output by AI which may initially appear to be believable, but which is factually, visually, sonically, or grammatically incorrect, or in which words, images or sounds are jumbled or do not make sense.
“Industry Standards” means those present or future standards, guidelines, protocols, and policies applicable to, governing, or addressing AI, including ISO 23894 and the National Institute of Standards and Technology AI Risk Management Framework, and any similar standards that have generally been recognized in the industry as establishing or promoting AI best practices.
“Input” means any data, Content, or information, submitted to an AI System or other AI service, deliverable or feature.
“Memorization” means when AI reproduces (in whole or in part) as an output any material inputs, prompts or other Content that was used to train, retrain, prompt, fine tune or otherwise develop such AI.
“Output” means any data, Content, or information generated by an AI System or other AI service, deliverable or feature.
“Training Data” means any Client Data that is used, sourced, or otherwise Processed in connection with this Agreement in connection with the development or training of any AI System or other application of AI under the Agreement.