AI Addendum

Scope. This AI Addendum forms part of the Agreement, which means the MuchSkills Enterprise Terms and the documents they incorporate, published at muchskills.com/legal. It applies to AI-assisted features of the platform, including any feature that generates a document about an identifiable individual and any AI-assisted feature MuchSkills adds after the date of this addendum. The intended purpose of the AI features is to generate content and suggestions for review by a person in connection with skills management, development and workforce planning. A plain-English summary of how MuchSkills uses AI is published as Our Approach to AI; this addendum, not that summary, states the binding terms. The AI providers in use, and the location of processing, are on the sub-processor list; changes are notified under the Data Processing Agreement.

1.  Roles and classification

1.1 MuchSkills is the provider of the AI features. The Customer is the deployer (or the equivalent under applicable AI law) when it uses them, including for any employment-related purpose. Each party complies with the AI law that applies to it in its role.

1.2 MuchSkills assesses the classification of its AI features under applicable AI law before introducing or materially changing a feature. Where a feature is or becomes a high-risk AI system, MuchSkills meets the provider obligations that apply to it, offers it under a separate written schedule allocating provider and deployer responsibilities, or withdraws or modifies it, and informs the Customer of any deployer obligation that then applies to the Customer. This addendum does not shift to the Customer a provider obligation that the law places on MuchSkills; the Customer's obligations as deployer are additional, not a substitute.

2.  Decision-support only

2.1 The AI features are a decision-support tool. They produce content, for example skill descriptions, development plans, summaries and generated CV content, for a person to review. They do not make automated decisions producing legal or similarly significant effects within Article 22 of the GDPR. Where an output relates to a named individual, the Customer may review, change or disregard it before use, and MuchSkills marks AI-generated content as provided in clause 4.1. No output is a recommendation as to a person's employment, and MuchSkills does not present it as one. The Customer remains the decision-maker for every employment-related decision, applies meaningful human review, and does not use an output as the sole or determinative basis for a decision about an individual. Where the Customer uses an output as a determinative input to such a decision, the Customer is responsible for the requirements of Article 22, and MuchSkills provides the information about the logic and significance of the processing that the Customer reasonably requires.

3.  Data handling

3.1 For most features, content is minimised before it leaves the platform so that it does not, on its own, identify an individual. Where a feature necessarily operates on an individual document, the content processed is personal data, handled under the Standard Contractual Clauses and the measures in the Data Processing Agreement for any transfer outside the European Economic Area.

3.2 MuchSkills does not use Customer Data to train or improve any third-party AI model, and its AI sub-processors are contractually prohibited from doing so.

4.  Transparency

4.1 MuchSkills informs users, where required, that they are interacting with an AI feature, and marks AI-generated content in a machine-readable form where required and technically feasible. The Customer is responsible for any notice it must give to its own workers and their representatives about use of the platform, and MuchSkills provides supporting information for that purpose.

5.  Prohibited use

5.1 The Customer must not, and must not permit a user to: (a) use the platform as the sole or determinative basis for a hiring, promotion, termination, redundancy or pay decision; (b) use it for automated decision-making within Article 22(1) of the GDPR, or for covert monitoring or surveillance of workers; (c) modify the models, scoring logic or thresholds, or feed outputs into a modified downstream model; or (d) rebrand or resell the platform under the Customer's own name, except under a separate written agreement allocating provider responsibilities.

6.  Change of purpose

6.1 MuchSkills specifies that the platform is not to be modified, rebranded or put to an intended purpose that would make it a high-risk AI system, or that would make the Customer a provider of it, under Regulation (EU) 2024/1689. If the Customer rebrands the platform, substantially modifies it, or changes its intended purpose so that it becomes a high-risk AI system, the Customer assumes the obligations of a provider under that Regulation and indemnifies MuchSkills against claims, fines and costs arising from that use. MuchSkills' own statutory duty to cooperate and provide information is unaffected, and MuchSkills provides documentation and technical assistance for that purpose at its then-current rates.

7.  Documentation

7.1 MuchSkills makes available, and keeps reasonably current, a model and system fact sheet describing the intended uses, known inappropriate uses, the categories of data used, known limitations and guidance on human oversight; the Annex to this addendum is that fact sheet. MuchSkills makes available logs of the Customer's use of the AI features, in the format the platform generates and for the retention period stated in Annex 2 to the Data Processing Agreement. MuchSkills makes available guidance and training materials on the appropriate use of the AI features.

8.  Outputs

8.1 AI outputs may be inaccurate or incomplete, and the Customer must review them before relying on them. MuchSkills does not warrant the accuracy, completeness, fairness or fitness of any output, or that the Customer's use complies with employment or equality law; that responsibility is the Customer's.

Annex  Model and system fact sheet

Intended uses. The AI features assist with skills work, for example generating skills content and descriptions, development plans and summaries, and generating documents such as curricula vitae, together with indicators about an individual derived from that content. The same applies to similar AI-assisted features that MuchSkills may add over time. Outputs are content and indicators for a person to review. An indicator about an individual, such as an indicative level, can be changed or set aside in the product by the person responsible, and it neither makes nor determines any decision about that individual.

Known inappropriate uses. As the sole or determinative basis for a hiring, promotion, termination, redundancy or pay decision; for automated decisions producing legal or similarly significant effects; or for covert monitoring of workers.

Categories of data used. For most features, generalised, non-identifying content; for any feature that necessarily operates on a document about an identifiable individual, curriculum-vitae content, which is personal data. No special-category data is intended; the Customer controls what it enters.

How it works, in outline. The features use third-party generative AI models to produce text or suggestions from the input described. The providers in use, and the location of processing for each, are published on the sub-processor list and kept current there. MuchSkills does not train those models on Customer Data, and the providers are contractually barred from doing so.

Known limitations. Outputs can be inaccurate, incomplete or generic, and can reflect biases in language or input data; they do not verify facts or predict future performance; quality depends on the input data the Customer provides.

Human oversight. Outputs are presented to a person, who can accept, change or set them aside and who makes any decision. Customers should keep a human in the loop for any decision affecting an individual and monitor for unfair or unexpected results. MuchSkills provides log access to support oversight.