Artificial Intelligence Addendum
This Artificial Intelligence Addendum (“AI Addendum”) supplements the Agreement between ACA and Client and applies to the extent AI Technology and/or AI Features are used in connection with the Services or System. Capitalized terms not defined in this AI Addendum have the meanings given to them in the Agreement.
Except as specifically related to AI Technology and AI Features, this AI Addendum does not modify the terms of the Agreement. In the event of a conflict between this AI Addendum and the Agreement solely with respect to the use of AI Technology and AI Features in connection with the Services or System, this AI Addendum will control. In all other respects, the Agreement shall remain in full force and effect.
1. Definitions.
“ACA Proprietary Information” means ACA’s Proprietary Information as defined in the Agreement as well as internal drafts, workpapers, notes, prompts, or other proprietary or internal materials.
“Agentic AI” means AI technology (developed by ACA or a third party) capable of autonomously executing multi-step tasks, making sequential decisions, or taking actions with limited human or no intervention, including where such functionality operates through the use of tools, APIs, or external data sources, executes instructions across multiple steps or iterations to achieve a defined objective.
“AI Features” means ACA developed prompts and software features and functionality made available directly or indirectly to Client or its Authorized Users (or ACA personnel on their behalf) through the Services or System that connect to AI Technology. For the avoidance of doubt, AI Features shall be included in the definition of “System.”
“AI Technology” means artificial intelligence, machine learning, natural language processing, automatic speech recognition, optical character recognition, large language models, generative AI, Agentic AI, or similar technologies, including technologies provided by third parties. For the avoidance of doubt, AI Technology shall not be included in the definition of “System.”
“AI Input” means data, content, materials, or information submitted to AI Features by Client or its Authorized Users, including where ACA personnel operate AI Features on behalf of Client, or its Authorized Users, under a managed services engagement or otherwise.
“AI Output” means output generated by AI Features based on AI Input and made available by or on behalf of Client or its Authorized Users.
2. ACA AI-Assisted Services.
ACA may use AI Technology to assist in performing the Services, developing and supporting the System, preparing work product, analyzing materials, developing drafts, performing quality control activities, benchmarking, or otherwise supporting ACA’s performance under the Agreement.
ACA’s use of AI Technology and/or Client’s use of AI Features will not cause ACA’s Proprietary Information to become AI Output, Client Data, or Deliverables solely as a result of AI Technology and/or AI Features being used by ACA personnel on behalf of Client.
3. Client Direct / Indirect Access AI Features, AI Input, and AI Output.
Where AI Input is submitted to, or AI Output is received or accessed from, AI Features by Client, its Authorized Users, or ACA personnel operating AI Features on behalf of Client or its Authorized Users under a managed services engagement, such AI Input and AI Output, excluding ACA Proprietary Information and any public information, will be treated as Client Data under the Agreement.
Except as otherwise expressly set forth in the Agreement, nothing in this AI Addendum grants Client any right, title, or interest in or to the System, AI Technology, AI Features, or ACA Proprietary Information.
For the avoidance of doubt, Client’s ownership of AI Output does not transfer or assign any right, title, or interest in ACA Proprietary Information, whether or not (i) used to generate, (ii) reflected in, (iii) incorporated into, or (iv) necessary to use, such AI Output. To the extent any AI Output incorporates or depends on ACA Proprietary Information, ACA grants Client a non-exclusive, non-transferable license to use such ACA Proprietary Information solely as embodied in the AI Output and solely for Client’s internal business purposes.
4. No Model Training and Data Processing.
Unless Client otherwise explicitly consents in writing, ACA does not use Client Data to train AI models and prohibits its third-party service providers of AI Technology from using Client Data to train their AI models.
Client Data may be processed by ACA or its applicable third-party service providers, including through the use of applicable AI Technology, as permitted pursuant to the Agreement and as necessary to provide, support, operate, and maintain the security of, the Services or System.
5. AI Outputs and Service Scope; Disclaimers.
Except as expressly set forth in this AI Addendum, ACA’s use of AI Technology and the use of AI Features by or on behalf of Client or its Authorized Users does not alter the (i) agreed upon Services, (ii) nature and treatment of Deliverables, or (iii) the parties’ respective rights and obligations under the Agreement, including any applicable EL, Statement of Work, Order Form, or other ordering document between Client and ACA (“Service Agreement(s)”).
Materials, analysis, or other information generated or developed through the use of AI Technology and/or AI Features are not intended to independently constitute any final decision or determination. ACA and Client’s use of AI Technology and/or AI Features does not guarantee that the Services, or the results thereof, will be favorably received by any regulatory agency or governing body.
Client acknowledges that AI Output may contain inaccuracies, errors, omissions, unsupported statements, or fabricated information and in each case should be reviewed and validated before use. ACA does not warrant that AI Output will be accurate, complete, current, error-free, non-misleading, or suitable for any particular purpose. Client is responsible for determining whether and how to use AI Output, and shall comply with all obligations and commitments in the Agreement and this AI Addendum in connection with Client’s and its Authorized Users’ use of the AI Features.
Client and its Authorized Users are solely responsible for the AI Inputs and the use of any AI Outputs, and any Client-configured settings, parameters, rules, thresholds, or other configurations applicable to the AI Features (“Client Modifications”). Notwithstanding anything to the contrary in the Agreement, and except to the extent that AI Output consists of ACA Proprietary Information, ACA shall not be liable, and shall have no indemnification obligation, for any claim, loss, damage, cost, expense, or liability arising from or
relating to AI Output or Client Modifications, including, but not limited to, Client’s or any Authorized User’s use, modification, distribution, publication, reliance on, or other exploitation of AI Output, and including any claim that AI Output or Client Modifications infringe, misappropriate, or otherwise violate any copyright, patent, trademark, intellectual property, proprietary, privacy, publicity, or other third-party rights.
6. Confidentiality, Security, and Proprietary Information.
Nothing in this AI Addendum limits or modifies either party’s confidentiality, privacy, data protection, and information security obligations applicable to such party under the Agreement.
Except to the extent expressly agreed in writing between the parties, nothing in this AI Addendum shall be deemed to require ACA to disclose ACA Proprietary Information, including any source code, prompts, prompt libraries, model orchestration logic, configurations, detection methodology, or other proprietary technical or operational details.
7. Use Restrictions and Capacity Controls.
Client shall not use, or permit any Authorized User to use, the AI Features in a manner that is inconsistent with the intended use of the applicable Services or System, including to process workloads, data, or requests for third parties, operate a separate or competing service, circumvent usage or access controls, or engage in automated, high-volume, excessive, or abusive usage that may adversely affect the performance, security, availability, cost, or integrity of the Services, System, or AI Features.
Client shall not, and shall not permit any Authorized User to: (i) use the AI Features to train, develop, fine tune, benchmark, or otherwise improve any competing AI model or service, including any proprietary model; (ii) attempt to extract, reverse-engineer, or replicate the underlying models, weights, or prompts of any AI Feature or AI Technology; (iii) engage in prompt injection, jailbreaking, or other attempts to circumvent the intended operation or safety controls of any AI Feature or AI Technology; (iv) use AI Outputs to evade ACA’s surveillance, audit, or compliance functions; (v) submit as AI Input to any AI Feature any data that such feature is not designed or authorized to handle, including certain categories of sensitive data (such as protected health information, material non-public information, or similarly regulated information), unless Client has received explicit, feature-level authorization from ACA; or (vi) access the AI Features through automated or programmatic means beyond ACA’s documented application programming interfaces.
Without limiting ACA’s other rights under the Agreement, ACA may apply reasonable usage limits, rate limits, capacity controls, or other technical restrictions, or suspend or limit access to the applicable AI Features, where ACA reasonably determines that such action is necessary to protect the Services, System, AI Features, AI Technology, ACA, its vendors, or other clients, or to address usage that exceeds applicable limits or is otherwise inconsistent with the Agreement.
For clarity, any issue relating to the accuracy, completeness, or quality of any materials or information generated or developed through the use of AI Technology, standing alone, will not constitute unavailability, material deficiency, downtime, or a service availability failure under the Agreement. Client shall not provide any AI Inputs or otherwise use the AI Features or AI Output in a manner that: (i) infringes or misappropriates any third party’s intellectual property rights or other proprietary rights; (ii) is deceptive, discriminatory, biased, unethical, defamatory, or illegal; or (iii) contains or introduces any viruses, worms, or other malicious computer programming codes that may damage the Systems. ACA reserves the right to suspend or terminate Client’s access to the Services for any failure by Client to comply with the requirements of this AI Addendum.
8. Cooperation.
Upon Client’s reasonable request, ACA shall provide timely cooperation and information relating to the AI Features as necessary for Client to meet its legal and regulatory obligations arising from the use of AI Features. Such cooperation may include providing reasonable details regarding the AI Features, including safety and quality measures, data sources and data handling practices, and any applicable third-party providers involved in the delivery of the AI Features. ACA’s obligations under this Section 8 shall be limited
to information within ACA’s reasonable knowledge and control and shall not require ACA to disclose information that is confidential, proprietary, or subject to third-party restrictions. Client shall provide reasonable advance notice of any such request and shall limit requests to information reasonably necessary for the stated compliance purpose.
9. Governing Law.
This AI Addendum and all matters arising out of or relating to this AI Addendum, whether sounding in contract, tort, or statute are governed by, and construed in accordance with the laws of the State of New York, United States of America (including its statutes of limitations and N.Y. Gen. Oblig. Law § 5-1401), without giving effect to the conflict of laws provisions thereof to the extent such principles or rules would require or permit the application of the laws of any jurisdiction other than those of the State of New York.