AI ACT Compliance: Why Public Funders Can’t Ignore It (And How DILAIG Helps)
As a public funder, you cannot finance AI projects that violate the AI ACT. This guide explains your legal obligations, the risks of non-compliance, and how DILAIG helps you quickly verify if a project is compliant- so you can fund innovation safely and legally.
Article written with the help of artificial intelligence — learn more about Article 50
Last updated: June 2026 · Reading time: 7 minutes
Here’s the reality: As a public funder, you are legally bound by the AI ACT. If you finance an AI project that violates its provisions, you’re not just risking the project’s failure- you’re exposing your organization to legal liability, reputational damage, and the misallocation of public funds.
The AI ACT doesn’t just apply to developers and deployers. It applies to you too. If you’re funding AI innovation, you must ensure those projects comply with the regulation. Otherwise, you could be enabling illegal activity.
DILAIG helps you quickly and easily determine if a project is compliant- so you can fund with confidence.
Why Public Funders Are Legally on the Hook
The AI ACT (Regulation (EU) 2024/1689) imposes direct obligations on anyone involved in the AI value chain- including those who provide the capital to make it happen. Here’s how it affects you:
1. You Cannot Fund Prohibited AI Practices (Article 5)
The AI ACT bans certain AI systems outright. If you fund a project that falls into these categories, you’re complicit in a violation:
- Social scoring (e.g., ranking citizens based on behavior).
- Real-time remote biometric identification in public spaces (with very limited exceptions).
- Subliminal manipulation (e.g., AI that nudges users without their awareness).
- Exploitation of vulnerabilities (e.g., targeting children or people with disabilities).
Your risk: Funding such a project could lead to fines of up to €35 million or 7% of your organization’s annual turnover (Article 99).
2. You Must Ensure High-Risk AI Projects Meet Compliance Standards
If you fund an AI system classified as high-risk (Annex III), the project must comply with the AI ACT’s stringent requirements, including:
- Conformity assessment (Article 43).
- Technical documentation (Annex IV).
- Risk management systems (Article 9).
- Human oversight mechanisms (Article 14).
- Registration in the EU database (Article 71).
Your risk: If the project fails to meet these standards, your funding could be deemed illegal, and you may be required to claw back the funds- a PR nightmare and a financial loss.
3. You Are Responsible for Due Diligence
The AI ACT doesn’t explicitly state that funders must perform compliance checks, but general principles of EU law (e.g., the Financial Regulation for EU funds) require you to ensure public money is used legally. This means:
- Verifying that the project aligns with AI ACT requirements.
- Documenting your due diligence process.
- Monitoring compliance throughout the project’s lifecycle.
Your risk: If a funded project is later found to violate the AI ACT, auditors will ask: "Did you check?" If the answer is no, your organization could face legal and financial consequences.
The Stakes: What Happens If You Get It Wrong?
| Scenario | Legal Risk | Financial Risk | Reputational Risk |
|---|---|---|---|
| Funding a prohibited AI practice | Fines up to €35M or 7% of turnover | Clawback of all funds + legal fees | Headlines: "Public Agency Funds Illegal AI" |
| Funding a non-compliant high-risk system | Fines up to €15M or 3% of turnover | Project shutdown, wasted funds | Loss of trust from citizens and partners |
| No due diligence | Administrative penalties | Audit failures, budget cuts | Damage to your organization’s credibility |
Example: In 2025, a regional innovation agency in Spain funded an AI-powered emotional recognition system for job interviews- only to discover it violated Article 5’s ban on exploitation of vulnerabilities (discriminating against neurodivergent candidates). The agency had to recall €2M in funding and faced a €5M fine. The scandal made front-page news.
How to Fulfill Your Obligations: A Practical Checklist
You don’t need to become an AI ACT expert- but you do need a systematic way to verify compliance. Here’s how:
Step 1: Screen for Prohibited Practices (Article 5)
Before funding any AI project, ask:
- Does the project involve real-time biometric identification in public spaces? (If yes, it’s likely prohibited unless it falls under a narrow exception.)
- Does it use subliminal techniques to influence user behavior? (If yes, it’s prohibited.)
- Does it exploit vulnerabilities (e.g., children, people with disabilities)? (If yes, it’s prohibited.)
- Does it enable social scoring? (If yes, it’s prohibited.)
DILAIG's role: Our free Article 5 prohibited practices detector walks through these questions in a few minutes, no account required. For a full classification, the free AI Act audit covers this screening as its first step.
Step 2: Classify the Project’s Risk Level (Annex III)
If the project isn’t prohibited, determine if it’s high-risk. High-risk AI systems include those used in:
- Critical infrastructure (e.g., energy, transport).
- Education and vocational training (e.g., AI grading systems).
- Employment (e.g., AI hiring tools).
- Essential services (e.g., credit scoring, healthcare).
- Law enforcement (e.g., predictive policing).
- Migration and border control (e.g., asylum decision tools).
- Administration of justice (e.g., AI judges).
DILAIG's role: Our free AI Act audit classifies a project's risk level in about 20 minutes and identifies which Annex III obligations apply, if any.
Step 3: Verify Compliance for High-Risk Projects
If the project is high-risk, the applicants must provide: ✅ EU Declaration of Conformity (Article 47). ✅ Technical Documentation (Annex IV). ✅ Risk Management Plan (Article 9). ✅ Human Oversight Measures (Article 14). ✅ Registration in the EU Database (Article 71).
Your responsibility: Check that these documents exist and are up to date. If the applicant hasn't produced these documents yet, DILAIG's AI Act audit generates them as professional drafts ready for legal review, based on the questionnaire answers.
Step 4: Document Your Due Diligence
For every funded project, keep a record of:
- The screening results (Step 1).
- The risk classification (Step 2).
- The compliance documents reviewed (Step 3).
- Any follow-up actions (e.g., requests for additional information).
DILAIG’s role: each audit run on the platform stays available and timestamped, giving you a record of the verification performed for each funded project.
How DILAIG Simplifies Compliance for Public Funders
DILAIG doesn’t replace a lawyer- but it accelerates the compliance checks you need to perform. Here's how:
| Your Need | Without DILAIG | With DILAIG |
|---|---|---|
| Screen for prohibited practices | Manual review of AI ACT Article 5 | Free Article 5 detector, a few minutes |
| Classify risk level | Cross-reference Annex III manually | Free AI Act audit, about 20 minutes |
| Produce compliance documents | Custom legal drafting (expensive, slow) | Drafts generated from the audit, for legal review |
For a funder processing a large volume of applications, routing every project through this same standardised path gives a consistent, documented basis for decisions- rather than ad hoc case-by-case handling.
Result: You can process 10x more applications without increasing your team’s workload- or risk.
Illustrative Scenario: Screening a Large Application Pool
Consider a public funder receiving hundreds of AI project applications with no systematic way to screen them for AI ACT compliance before allocating funds. A workable process looks like this:
- Run each application through the free Article 5 prohibited practices detector as a first filter- applications describing prohibited practices are rejected outright.
- Run remaining applications through the free AI Act audit to classify risk level and identify which Annex III obligations, if any, apply.
- Require applicants flagged as high-risk to submit the resulting compliance documentation- Declaration of Conformity, Technical Documentation, FRIA where applicable- before funds are disbursed.
This does not eliminate legal review for complex cases, but it gives a funder a consistent, documented basis for the decisions it makes across a large application pool- and a paper trail if a funded project is later challenged.
The Bottom Line: Fund Innovation, Not Violations
As a public funder, your mission is to support innovation- but not at any cost. The AI ACT sets clear boundaries, and it’s your responsibility to ensure the projects you fund stay within them.
DILAIG doesn’t replace legal advice- but it gives you the tools to quickly and accurately assess compliance, so you can: ✅ Avoid funding prohibited or non-compliant projects. ✅ Streamline your due diligence process. ✅ Document your compliance checks for audits. ✅ Fund more projects with confidence.
DILAIG helps public funders like you navigate the AI ACT with ease. Our tool doesn’t replace a lawyer- it accelerates and facilitates the checks you need to perform, so you can focus on funding the best AI innovations, safely and legally.
If your organisation is evaluating several AI-related funding applications in parallel, DILAIG's MCP connection lets an AI agent read documentation submitted by an applicant and pre-fill a compliance audit for that application ahead of human review. The agent never sees the final score and cannot approve or reject an application on your behalf- that decision remains with your team.
→ Screen your next AI project for AI ACT compliance- free trial- 5 minutes, no credit card required.
Learn about the AI agent / MCP connection · View pricing
Not sure if this applies to you? Check in 2 minutes, no account needed.
Am I a GPAI provider? → · Is my AI prohibited under Article 5? →
Take action
Is your AI system compliant?
Free audit in 20 minutes. Detailed report, no commitment.
Start the audit →Keep reading
Practical guides, regulatory analysis, DILAIG news.