This AI Addendum (the “Addendum”) is entered into and effective as of {DATE} (the "Effective Date"), by and between the Parties to the Master SaaS Agreement (the “Agreement”) entered into as of {DATE} and forms an integral part thereof including also the Data Processing Agreement (the “DPA”) Capitalized terms used but not defined in this Addendum shall have the meaning set forth in the Agreement.
1. Purpose and Scope
(a) This Addendum governs the use of artificial intelligence and machine learning features (“AI Functionality”) made available by Starred as part of the Services under the Agreement.
(b) The AI Functionality is designed solely to support feedback analysis and experience optimization in the recruitment process. The Services, including any AI Functionality, shall not be used to evaluate, select, or reject candidates or to make automated decisions towards individuals.
2. Definitions
(a) “AI Functionality” means any part of the Services involving algorithmic processing, including (a) machine learning models used to normalize feedback across geographies and candidate types (e.g., the “Standard Index” and “Priority Matrix”), and (b) generative AI or large language models used to analyze open-text feedback and generate insights.
(b) “Customer prompts” means prompts phrased, defined and submitted by the Customer (or its Authorized Users) to the AI Functionality.
(c) “GDPR” means the Regulation (EU) 2016/679 of the European Parliament and of the Council of 27 April 2016 on the protection of natural persons with regard to the processing of personal data and on the free movement of such data and repealing Directive 95/46/EC (General Data Protection Regulation).
(d) “Input” means any data, content, or information provided by the Customer to the System as part of a prompt to be used to define context for the AI Functionality.
(e) “Output” means data, content, or information generated by the AI Functionality and presented back to the Customer from AI Features based on Input provided to the AI Functionality by Customer. Outputs may derive from Customer Data but do not include Customer Data itself.
(f) “Personal data” shall have the meaning ascribed to it in the General Data Protection Regulation (GDPR).
(g) “Predefined Prompts” means prompts, instructions, queries, or input templates generated, configured, or embedded by Starred as part of the AI Functionality.
(h) “Third-Party Models” means any generative AI or LLM functionality provided by third-party vendors (e.g., OpenAI) optionally made available as Third-Party Integrations. The Third-Party Models currently in use by Starred are listed in Appendix A.
(i) “The Agent feature” means a prompt-driven component of the AI Functionality that operates exclusively in response to Customer Prompts and assists the Customer in understanding, exploring, and analysing Customer Data within the analytics features of the Services by generating explanations, summaries, or analytical insights.
3. Ownership
(a) Customer acknowledges that, as between the Parties, Starred and/or its licensors own all rights, title, and interest, including all intellectual property rights, in and to the AI Functionality.
(b) Starred acknowledges that, as between the Parties, the Customer exclusively owns all rights, title, and interest in and to all Input and Output, including any intellectual property rights therein. Starred makes no claim to any such rights. Notwithstanding the foregoing, Predefined Prompts shall remain the sole property of Starred.
(c) Customer grants Starred a non-exclusive, royalty-free license to use Input and Output as necessary to deliver the Services, including the AI Functionality, in accordance with the Agreement and this Addendum. This includes a license to use anonymized Customer Data for internal benchmarking and model improvement purposes, as expressly permitted under Section 4.
(d) Customer may not use AI Functionality to develop its own artificial intelligence products, services, features, or technology, or any machine learning models or technology.
4. Use of Customer Data for Training
(a) As an indispensable part of the AI Functionality, Starred will monitor and analyze Customer’s use, and collect and compile data and information related to such usage, for the sole purpose of enhancing the AI Functionality (and thereby the Services) and generating anonymized benchmarks that provide valuable comparisons and performance metrics, allowing customers to assess their own performance within a broader context (“Aggregated Statistics”), cf. Clause 5(d). The Aggregated Statistics do not include any (i) information identifying Customer or any individual or (ii) Customer’s Confidential Information.
(b) Starred shall not use Input or Output – including any possible Customer Data included therein – to train or fine-tune any Third-Party Models. For the avoidance of doubt, this restriction applies regardless of whether such models are hosted by Starred or accessed via third-party APIs.
(c) Provided Starred ensures that Input, Output and Customer Data cannot be linked to an identified or identifiable natural person or to Customer itself, Starred is permitted to use such anonymized data to train and calibrate internal machine learning models for the purposes of for quality assurance, statistical research purposes, benchmarking and improving prioritization models, subject to the data handling requirements in Section 6.
5. Prompt Use, Prompt Controls and User restrictions
(a) The AI Functionality includes Predefined Prompts and Customer Prompts, subject to the terms of this Addendum.
(b) Starred shall maintain commercially reasonable technical and organizational controls to ensure that the AI Functionality – including prompt functionality – operates in a secure and compliant manner. Depending on the use-case, such controls may include input validation, logging, prompt filtering, and output moderation, depending on the applicable use case.
(c) Starred shall use commercially reasonable efforts to prevent and remediate algorithmic bias across the AI Functionality.
(d) When defining and phrasing content for Customer Prompts, the Customer is solely responsible for ensuring that such content:
- (i) complies with applicable law,
- (ii) does not include personal, confidential or sensitive data, and
- (iii) is not used to solicit or generate unlawful, discriminatory, deceptive, misleading or otherwise harmful content.
(e) The Customer is further responsible for verifying the suitability of any Output before relying on it for any purpose. Customer must not rely solely on any Output and is expected to apply human oversight in all use cases involving Output.
(f) Customer shall not use the AI Functionality:
- (i) To automate or influence hiring decisions.
- (ii) To mislead anyone that an Output was solely human generated.
- (iii) For high-risk or safety-critical activities (including but not limited to legal, employment, medical, or financial determinations).
- (iv) In violation of applicable law or regulation.
- (v) To develop, train, or reverse-engineer competitive AI models or extract any components of AI Functionality (such as using prompts to discover training data).
- (vi) To define and phrase Customer Prompts or otherwise provide Input to the AI Functionality with the purpose of generating offensive, discriminatory, or unlawful content.
(g) The Customer is solely responsible for developing own internal policies regarding the appropriate use of the AI Functionality and for training Authorized Users.
(h) Starred may log and retain Input and Customer Prompts together with corresponding Output for the purpose of maintaining commercially reasonable technical and organizational controls as set out in clause 5(b).
(i) Starred reserves the right to restrict, suspend, or condition access to prompt functionality in whole or in part if Starred has reason to suspect
- (i) misuse (or attempt to misuse) on Customer’s/Authorized User’s part,
- (ii) that Customer’s access poses a legal or compliance risk, or
- (iii) that Customer’s access may cause material harm to Starred, to Starred’s systems or to any third party.
6. Anonymization and Data Protection
(a) Starred shall ensure that the AI Functionality – when generating Output – processes only anonymized data – thus data that cannot be linked to an identified or identifiable natural person. Personal data shall be segregated and inaccessible to AI components through technical and organizational safeguards.
(b) Starred shall maintain appropriate technical and organizational measures to ensure separation of personal data and enforce anonymization “by design,”, including but not limited to access controls, prompt validation, and suppression of identifiers such as IP addresses and timestamps.
(c) Starred shall use commercially reasonable efforts to monitor and restrict input mechanisms, including through validation logic and interface design, to prevent the inadvertent submission of personal data in free-text or in any possible customized data field. Starred shall ensure such data is excluded from the AI Functionality. A blacklist-based validation system shall be in place to prevent unauthorized use of free-text fields for Personal Data.
(d) Nothing in this Addendum modifies or limits the Parties’ obligations under applicable data protection laws. The Parties acknowledge that all processing of personal data is governed by a separate Data Processing Agreement (DPA) entered into between the Parties.
7. Third-Party Models
(a) The Agent feature includes optional AI functionality that may be powered by Third-Party Models. The Third-Party Models available for use with the Agent feature are listed in Appendix A. Use of such models is optional and subject to Customer’s prior enablement.
(b) Where Third-Party Models are enabled, the Customer agrees to comply with the third-party provider’s applicable terms of service and acceptable use policies. Starred shall make such terms available upon request or within the Services.
(c) Starred shall maintain appropriate technical and organizational measures to ensure separation of personal data and enforce anonymization “by design,” including but not limited to access controls.
(d) Starred shall implement commercially reasonable safeguards to ensure that any Customer Data submitted to Third-Party Models is limited, anonymized, and transferred only to the extent necessary to provide the AI functionality as enabled by Customer. Starred shall also use commercially reasonable efforts to minimize and monitor potential bias in the configuration and application of such Third-Party Models, where technically feasible.
(e) If Third-Party Models are enabled, the Customer shall be responsible for reviewing and accepting the third-party vendor’s terms and acceptable use policies.
8. Disclaimers
(a) The AI Functionality and all related Output are provided “as is” and without warranty of any kind. Output that appears accurate because of its detail or specificity may still contain material inaccuracies. Starred makes no representation or warranty, express or implied, that Output will be accurate, complete, current, error-free, or suitable for any particular purpose.
(b) Output generated by the AI Functionality may not be unique and may be similar to Output generated for other customers. Starred does not guarantee the originality or exclusivity of any Output.
(c) Starred likewise disclaims all liability for the quality, reliability, or legality of any Output generated by Third-Party Models. The Customer acknowledges that Starred does not control how such models generate Output and shall not be liable for any inaccuracies, bias, or violations arising therefrom.
9. AI-Transparency and Information
(a) Starred shall upon Customer’s reasonable written request use commercially reasonable efforts to provide transparency and clarifications relating to the AI Functionality to the extent necessary for Customer’s compliance purposes
(b) Starred shall upon such request present Customer with documentation and information enabling the Customer to reasonably,
- (i) understand the nature, intended purpose and main characteristics of the AI Functionality,
- (ii) use the AI Functionality appropriately and in accordance with the Agreement, and
- (iii) meet any transparency and information obligations applicable to Customer under applicable AI laws and regulations (including, where applicable), EU AI Act (Regulation 2024/1689 of 13 June 2024 on artificial intelligence)
(c) Depending on the nature and purpose of the Customer’s request, such documentation may include: (a) instructions for use; (b) descriptions of key capabilities and material limitations; (c) information on commercially reasonable efforts to monitor and restrict input mechanisms; and (d) information on reasonably foreseeable risks of inaccurate, incomplete or misleading Output and recommended human oversight.
(d) For the avoidance of doubt the obligations following from this Clause 9 shall not include an obligation for Starred to disclose source code, model weights, or other information that would reasonably be expected to compromise Starred’s security, intellectual property, or trade secrets.
(e) Any documentation or information provided under this Clause 9 constitutes Starred’s Confidential Information and may be used by Customer solely for internal compliance and governance purposes in connection with the use of the Services.
(f) For the avoidance of doubt, this clause does not modify or expand the Parties’ rights and obligations under the applicable DPA entered into between the parties as part of the Agreement. Any audit, inspection or access rights relating to the processing of personal data shall be governed exclusively by the DPA.
(g) Where AI Functionality involves Third-Party Models, Starred shall make commercially reasonable efforts to pass through or provide access to equivalent documentation from such Third-Party Models to the extent available and contractually permitted.
10. Indemnity and Liability
(a) The limitations of liability set forth in Section 10 of the Agreement shall apply equally to this Addendum.
(b) The Customer shall indemnify and hold harmless Starred from any Third-Party Claim arising from
- (i) the Customer’s misuse of the AI Functionality;
- (ii) inputting prohibited content; or
- (iii) breach of applicable use restrictions under this Addendum.
11. Precedence and Survival
(a) In the event of any inconsistency or conflict between the Agreement and this Addendum, this Addendum shall prevail although solely with respect to the AI Functionality and any matter expressly governed by this Addendum.
(b) Sections 3, 4, 5, 6, 8, 9 and 10 of this Addendum will survive any expiration or termination of the Agreement for so long as necessary to give the effect.
IN WITNESS WHEREOF, the Parties hereto have executed this AI-Addendum as of the Effective Date.
STARRED GROUP A/S | {CUSTOMER NAME}
Starred signature: ____________________ Customer signature: ____________________
Starred signer: ____________________ Customer signer: ____________________
Starred Title: ____________________ Customer Title: ____________________
Exhibit A: AI Model Inventory