AI Processing Notice

1. Purpose of this notice

1.1. This AI Processing Notice / AI Disclaimer explains how Grant Fund uses artificial intelligence, machine learning, large language models, retrieval systems, classification systems, and related automation in connection with the Service.

1.2. This Notice supplements the Terms and Conditions, Privacy Policy, Data Processing Addendum, and other applicable Grant Fund legal documents.

2. What AI is used for

2.1. Grant Fund may use AI-assisted systems for classification, extraction, source matching, summarisation, programme screening, readiness analysis, report generation, risk warnings, eligibility labels, negative-match memos, radar updates, gap analysis, source-reference generation, quality control, and workflow support.

2.2. AI may be combined with rule-based processing, deterministic screening, retrieval systems, internal classification logic, public-source matching, Official Source analysis, and human review workflows.

3. What AI is not used for

3.1. AI Outputs are not official Grantor decisions.

3.2. AI Outputs are not legal, tax, accounting, financial, investment, immigration, procurement, state-aid, sanctions, export-control, patent, intellectual-property, compliance, grantor, or regulated professional advice.

3.3. The Service is not designed or authorised for use as a high-risk AI system, credit-scoring system, employment decision system, migration decision system, public-benefit decision system, law-enforcement system, biometric-identification system, medical device, safety component, or automated decision system producing legal or similarly significant effects on individuals.

3.4. Users must not use the Service or Outputs to make high-risk automated decisions about individuals unless they have independently ensured full compliance with applicable law and obtained Grant Fund's prior written consent where required.

4. No official Grantor decision

4.1. Admissibility, eligibility, scoring, ranking, approval, rejection, award, payment, compliance, and interpretation are determined only by the relevant Grantor, funding body, agency, evaluator, evaluation panel, court, public authority, bank, investor, accelerator, programme operator, or other competent body.

4.2. Labels such as "fit", "eligible", "not yet", "not fit", "filtered", "ready", "gap", "recommended", "priority", "negative match", "deadline", "funding amount", and "programme status" are internal platform classifications only.

5. Error, hallucination, and limitation risks

5.1. AI Outputs may be inaccurate, incomplete, outdated, inconsistent, non-exhaustive, incorrectly classified, misleading, mistranslated, overconfident, non-traceable in full, unsuitable for a User's specific circumstances, or otherwise wrong.

5.2. AI systems may hallucinate, omit relevant information, misread Official Sources, misread annexes, misunderstand exceptions, mistranslate requirements, misclassify programmes, infer incorrectly, miss context, fail to capture unwritten interpretations, fail to capture updated source materials, or fail to identify conflicting rules.

5.3. The Service may fail to identify a relevant Funding Program or may include a Funding Program that is not suitable for the User.

6. Source traceability limits

6.1. Grant Fund may provide links, snippets, citations, source references, extracted conditions, or traceability indicators.

6.2. Traceability to Official Sources or source snippets does not mean full explainability of the AI model, completeness of analysis, legal correctness, official confirmation, or suitability for reliance.

6.3. Source references may be partial, delayed, unavailable, outdated, superseded, mistranslated, or affected by changes in Official Sources.

7. Human review limits

7.1. Some Outputs may be subject to human review where Grant Fund chooses or where a specific agreement provides for it.

7.2. Human review is not guaranteed unless expressly agreed in writing.

7.3. Human review does not convert an Output into legal advice, official eligibility confirmation, professional advice, or a guarantee.

8. User verification duty and no sole reliance

8.1. Users must independently verify all relevant Official Sources before acting, applying, submitting, signing, budgeting, hiring, engaging consultants, forming a consortium, contacting partners, or making strategic decisions.

8.2. Users must not use Outputs as the sole basis for legal, financial, investment, tax, grant, compliance, immigration, corporate, fundraising, hiring, product, employment, public-benefit, credit, insurance, procurement, sanctions, export-control, patent, intellectual-property, or strategic decisions.

8.3. Users must obtain independent professional advice where professional advice is required or prudent.

9. AI providers and data use

9.1. Grant Fund may use OpenAI, Anthropic, Mistral, self-hosted models, proprietary models, third-party AI vendors, infrastructure providers, retrieval providers, monitoring providers, and other AI or infrastructure providers.

9.2. Grant Fund does not share User Passport data with external AI providers unless disclosed in the Privacy Policy, DPA, or Subprocessor List.

9.3. User Data is not used to train third-party foundation models unless expressly disclosed and legally permitted.

10. AI logs and retention

10.1. Grant Fund may retain prompts, source excerpts, model inputs, model outputs, review logs, confidence metadata, audit logs, system traces, security logs, and related records where necessary or useful for safety, auditability, debugging, legal defence, quality control, abuse prevention, compliance, fraud prevention, security, product improvement, or performance monitoring.

10.2. Retention is governed by the Privacy Policy, Data Retention Summary, DPA where applicable, and Applicable Law.

11. Correction, withdrawal, regeneration, and reclassification

11.1. Grant Fund may correct, modify, withdraw, flag, suppress, reclassify, replace, regenerate, disable, or remove any AI Output, report, recommendation, score, memo, radar update, source reference, confidence indicator, label, or verdict at any time.

11.2. Users should promptly notify Grant Fund of suspected errors, outdated sources, incorrect labels, or safety issues.

12. AI Act positioning

12.1. Grant Fund is designed as a SaaS information, decision-support, screening, and workflow-support tool for grant eligibility analysis and funding-program workflow support.

12.2. Grant Fund is not designed as a high-risk AI system for credit scoring, employment, worker management, education access, migration, asylum, border control, law enforcement, administration of justice, democratic processes, public-benefit eligibility, medical devices, biometric identification, or safety components.

12.3. Users must not repurpose the Service into such systems or use Outputs in such contexts without independent legal review and Grant Fund's prior written consent where required.