💰 Funding Details

Leveraging Artificial Intelligence for Pandemic Preparedness and Response (HORIZON-HLTH-2025-01-DISEASE-04)

Key Facts

* Type of Action: HORIZON-RIA (Research & Innovation Action)

* Maximum EU Contribution per Project: €80 million (100 % of eligible direct costs + 25 % indirect costs)

* Opening Date: 22 May 2025

* Deadline: 16 September 2025, 17:00 Brussels time (single stage)

* Programme / Cluster: Horizon Europe – Cluster 1 *Health* 2025 Work Programme, Destination *“Tackling diseases and reducing disease burden”*

* TRL at start/end: Typically TRL 2-5 → TRL 5-7 (research-driven but with strong validation elements)

What Is Funded?

The call finances multi-disciplinary R&I projects that create or enhance trustworthy, ethical and FAIR AI tools spanning the full pandemic management cycle:

1. Prediction & Early Warning – e.g. real-time forecasting, wastewater or air biosurveillance, anomaly detection across heterogeneous datasets.

2. Prevention & Containment – AI-supported design of medical/non-medical countermeasures, logistics optimisation, behavioural intervention modelling.

3. Diagnosis & Clinical Management – differential diagnosis, risk stratification, treatment response prediction.

4. System Resilience – resource allocation, surge capacity modelling, cross-border data sharing architectures.

Projects must demonstrate EU value-added by harnessing transnational FAIR datasets, linking to infrastructures such as EHDS and EOSC, and by integrating SSH expertise to anticipate societal, ethical, gender and bias issues.

Budget Structure (Typical)

* Personnel & Subcontracting: 45-55 %

* Data acquisition/curation: 10-20 %

* Computing & AI development: 15-25 %

* Pilots/validation (incl. clinical studies): 10-15 %

* Clustering & dissemination: ≥2 % (mandatory inter-project collaboration)

* Contingency & ethics compliance: 3-5 %

> Tip: Allocate a clear work package for joint clustering activities and reserve travel + event costs from the outset.

📊 Evaluation Criteria

How Your Proposal Will Be Scored

Excellence (Weight 1/3, Threshold 4/5)

* Soundness of AI approach – novelty of algorithms, generative components, bias mitigation.

* Interdisciplinarity & SSH integration – meaningful roles, not tokenistic.

* Quality of FAIR data strategy – clear provenance, metadata, governance.

Impact (Weight 1/3, Threshold 4/5)

* Credible pathway to Destination impacts – quantified KPIs tied to disease-burden reduction.

* Measures to maximise uptake – standards, certification, open-source components, IP strategy for SMEs.

* Sustainability & exploitation – business models, links to your country’s health-tech clusters, policy endorsements.

Quality & Efficiency of Implementation (Weight 1/3, Threshold 4/5)

* Consortium capacity – complementary expertise, prior collaboration, gender balance in leadership.

* Work plan coherence & resources – Gantt chart, milestones, critical path.

* Risk management & ethics – GDPR compliance, cybersecurity, dual-use screening.

Only proposals scoring ≥12/15 overall and passing each threshold proceed to Grant Agreement.

📝 Application Strategy

Winning Strategy

1. Align with Expected Outcomes

Map every work package to the three required outcomes:

* Outcome 1 – Full AI potential unlocked → WP1 *Data federation & FAIRification*, WP2 *Generative AI model development*.

* Outcome 2 – Readily available, trustworthy tools → WP3 *Validation pilots in hospitals & PH agencies*, WP6 *Certification & CE-marking roadmap*.

* Outcome 3 – Multisource data integration → WP1 + WP4 *Interoperability & cross-border governance*.

2. Build a Balanced Consortium

* AI excellence: universities with strong ML portfolios.

* Domain datasets: public health institutes, hospitals, biobanks.

* Infrastructure operators: EOSC nodes, EHDS pilots.

* End-users: ministries, ECDC, regional PH authorities in your country & beyond.

* SSH & Ethics: experts in gender studies, AI ethics, health economics.

* SME innovators: start-ups developing edge AI, privacy-preserving analytics.

3. Demonstrate Impact Pathway

1. Short-term (0-3 yrs): validated AI prototypes deployed in at least three real-world test beds.

2. Medium-term (3-5 yrs): policy adoption by your country’s health authority; contribution to HERA stockpiles.

3. Long-term (5+ yrs): integration into EHDS routine workflows; commercial scale-up via SME partners.

Quantify KPIs (e.g. ≥15 % improvement in detection lag, ≥30 % reduction of manual data harmonisation effort).

4. Exploit Synergies

* Liaise with HORIZON-CL4-2021-HUMAN-01-24 projects on bias mitigation toolkits.

* Co-organise workshops with SC1-PHE-CORONAVIRUS-2020-2C behavioural projects.

* Connect to ERA4Health & EIT Health accelerators for exploitation.

5. Address Risks Proactively

* Data access risk: MoUs with data custodians signed before submission.

* Regulatory risk: early dialogue with EMA & national AI sandboxes in your country.

* Technical risk: redundancy in computational resources across 2+ HPC centres.

💡 Success Factors

Early Data Custodian Engagement

Secure **Letters of Support** from biobanks, hospitals and PH agencies before submission, detailing data volumes, modalities and legal clearances.

Living Lab Approach

Set up continuous co-design cycles with end-users (clinicians, policymakers, citizens) via **bi-monthly sprint reviews** to ensure usability and trust.

Bias-Audited Model Development

Integrate **model cards, bias impact statements and demographic parity tests** as formal deliverables, aligning with EU AI Act requirements.

Dual Exploitation Pathway

Combine **open-source core algorithms** (fostering uptake) with **proprietary value-added services** exploited by SME partners, ensuring both impact and revenue.

Regulatory Sandboxing

Leverage **AI regulatory sandboxes** in your country to pilot tools under real conditions, accelerating conformity assessment and CE-marking.

Clustering Leadership

Volunteer as **cluster coordinator** among funded projects, positioning your consortium to shape synergies, joint policy briefs and future calls.

❌ Common Mistakes to Avoid

Ignoring Cross-Border Data Governance

Proposers often assume national datasets can be shared EU-wide without legal mapping. **Provide a concrete plan** for legal bases (GDPR Art. 9(2)(i)), Data Governance Act compliance and Data Transfer Impact Assessments from your country to other Member States.

Token SSH Participation

Listing an ethics professor without allocated tasks fails to meet the *effective contribution* requirement. Assign SSH partners **leadership of a dedicated WP** on societal acceptance, behavioural insights and gender bias auditing.

Under-budgeting Joint Clustering Activities

EC expects collaboration across funded projects. Consortia that allocate **€0 travel or dissemination** for clustering are downgraded. Reserve ~2 % of budget and describe concrete joint deliverables.

Over-promising TRL Progression

Claiming TRL 9 within a 4-year RIA raises feasibility concerns. Align ambitions: **TRL 5-7** with pathway to regulatory approval post-project.

Neglecting Energy Footprint of AI

The call explicitly mentions energy consumption. Provide **carbon-aware training schedules, model compression strategies and green HPC sourcing** to avoid reviewer criticism.

Single-Country Pilots

Running validation only in your country undermines EU added value. Include **multi-site pilots** across ≥3 Member/Associated States.

Leveraging artificial intelligence for pandemic preparedness and response

Quick Facts for Spain Applicants

Ready to Apply from Spain?