Background

UNESCO has the mandate to promote the development of science and technology from an ethical and human rights approach. Under this framework in November 2021, UNESCO developed the Recommendation on the Ethics of Artificial Intelligence (AI), which the 193 Member States adopted by acclamation at the 41st session of its General Conference, and spearheading its implementation by the Member States through innovative tools and methodologies, such as the Readiness Assessment Methodology (RAM) and the Ethical Impact Assessment (EIA). Through the application of these tools, UNESCO is changing the operating model that drives AI, going beyond principles to develop a range of concrete and practical solutions to ensure fair, inclusive, sustainable and equitable outcomes of AI.

The Digital Flanders Agency has launched an Artificial Intelligence Expertise Center (AI EC) tasked with enhancing and guiding the adoption of AI across Flemish public administrations. The capacity and capability of the AI EC need to be strengthened to empower it to fulfill its coordination role among Flemish public administrations in the facilitation and implementation of AI projects.

To this end, UNESCO is implementing the AI-Ready Flemish Public Administration project, funded by the European Commission Directorate-General for Structural Reform Support (DG REFORM) through the EU's Technical Support Instrument (TSI). The project aims to foster Flemish public administrations' AI-readiness and increase the adoption of AI across the Flemish public administrations in a safe, robust, trustworthy, and ethical manner. Through its three pillars, the project focuses on strengthening: (i) institutional capacity for AI governance; (ii) technical capacity for AI projects; and (iii) human capacity for AI awareness, literacy, and skills within the Flemish public administrations.

The project operates through a close collaboration with Digital Flanders' AI Expertise Centre, whose mandate is to support the 90+ Flemish public administration entities through a 1-to-many approach. Flanders is also a world-first in implementing UNESCO's RAM at the sub-national level, mapping the AI policy and governance landscape and identifying priority areas for ethical AI governance across six dimensions: ethical governance, legal, socio-cultural, scientific, economic, and technical.

The Impact Assessment Priority

While international and EU-level frameworks, including the EU AI Act and UNESCO's own Ethical Impact Assessment methodology, provide important normative scaffolding, their practical application within sub-national public administrations requires contextualisation. Flemish public administration entities operate within a distinct legal and administrative context, with specific institutional capacities, regulatory obligations, and governance cultures that shape how impact assessment tools can and should function in practice.

The project has identified a pressing need for research-grounded, legally-informed guidance to support the AI Expertise Centre and its network of public administration entities in understanding, adapting, and applying impact assessment tools to their specific operational realities. This includes mapping the landscape of existing frameworks, identifying implementation gaps and barriers, and generating practical, accessible guidance to embed impact assessment as a routine element of AI governance in Flemish public administrations. It also includes working closely with the AI Expertise Centre to ensure that solutions respond to the fast changing policy environment.

Aim of this Contract for Services

The aim of this contract is to provide expert research and practical guidance on applied impact assessment tools for AI in the Flemish public administration context.

The contracted organisation will build on and integrate the learnings accumulated across previous phases of the AI-Ready Flemish Public Administration project to advance the practical operationalisation of impact assessment within the network of Flemish public administration entities. Rather than beginning from a blank-slate analysis of existing frameworks, the work is grounded in the concrete project experience to date, including working self assessment tools and a needs analysis for navigating the legal landscape.

The outcome is to translate these accumulated insights into accessible, applied outputs support the AI Expertise Centre and its network of public administration entities in operationalising impact assessment practices.

 

Overview of Functions

The contracted organization will work under the overall authority of the Assistant Director-General for the Social and Human Sciences Sector (ADG/SHS) and under the direct supervision of the Head of the Ethics of AI Team at UNESCO Headquarters. The contracted organisation will work closely with the Project Lead for content direction, quality feedback, and stakeholder coordination.

The outputs and activities the contracted expert will engage in are:

Output 1 – Literature Review

The contracted organization will conduct a structured review of the existing deliverables and literature produced under the AI-Ready Flemish Public Administration project, alongside a targeted analysis of relevant AI impact assessment frameworks at international, EU, and national levels. The aim is to ground subsequent outputs in the concrete experience and findings of the project, while situating these within the broader regulatory and governance landscape.

• While literature reviews and research have already been conducted, these must be updated to ensure they remain state of the art.
• Where available, the contracted organization shall draw on post-implementation reflections, stakeholder feedback, and monitoring data gathered by the AI Expertise Centre and project team. Lessons learned shall be treated as primary analytical material informing the scoping and recommendation work in Output 2, rather than presented only as background.

Output 2 - Stakeholder Engagement

The contract organization will design and implement a targeted stakeholder engagement process to identify the needs of Flemish public administrations with respect to the Impact Assessment framework design and constructions. This includes:

• Conducting online interviews with stakeholders to map needs, priorities, and existing practices related to Impact Assessments and AI governance challenges.
• Tentatively organising and facilitating a workshop in Brussels to gather structured feedback on the relevance and applicability of impact assessments

Output 3 - Blog Post

The contracted expert will produce one blog post translating key findings and insights from the impact assessment research into accessible content for a wider policy audience.

Topics will be selected in consultation with the Project Lead, drawing on the research findings and the practical experience of the Flemish project implementation. It may optionally include suggestions for infographics or supporting visuals, in line with UNESCO Communication Guidelines.

Output 4 – Guidance Report on Applied Impact Assessment

The contracted organization will produce a user-oriented Guidance Report on Applied Impact Assessment, designed to help Flemish civil servants and public administration entities find their way through the landscape of AI self-assessment tools and different use cases. The report is intended as a practical reference resource and decision-support framework for Flemish public sector practitioners, and may be developed as a training modality.

 

The Contractor shall address feedback provided by the Secretariat and incorporate it into all deliverables, as appropriate. In addition, the Contractor may be requested to contribute to the development of relevant communication and dissemination materials in support of the Secretariat’s objectives.

 

Deliverables and Timeframe

 

Deliverable

May 25, 2026 - Output 1: Review of Existing Project Literature and Analysis of Impact Assessment Frameworks

June 29, 2026 -  Output 2 and Output 3: Stakeholder engagement report and blog post

July 27, 2026 - Output 4: Navigation Report on AI Self-Assessment Tools (first draft), including minimum one round of review

 

Experience Requirements

Experience

• Minimum of ten years of professional experience in AI governance, legal analysis, or technology policy, with demonstrated focus on AI impact assessment frameworks or related instruments (such as AI risk assessment, fundamental rights impact assessment, or algorithmic accountability mechanisms).
• Proven experience conducting legal and/or policy analysis of EU digital regulation, including the EU AI Act, GDPR, and relevant Belgian or Flemish administrative law frameworks.
• Demonstrated experience producing research outputs, guidance documents, toolkits, or frameworks intended for a public administration or policymaker audience.
• Prior experience working with or advising public sector entities on AI governance, digital transformation, or responsible AI implementation is strongly desirable.

 

Skills and Competencies

• Strong command of the EU and international AI impact assessment landscape, including familiarity with UNESCO's Ethical Impact Assessment methodology, the EU AI Act's impact assessment requirements, and relevant sector-specific frameworks.
• Ability to translate complex legal and governance requirements into practical, accessible guidance for non-specialist practitioners in public administration.
• Proven capacity to produce high-quality analytical and practical outputs — including research reports, toolkits, and policy communications — within defined timelines.
• Excellent project management and coordination skills, with the ability to manage multiple interrelated deliverables and maintain productive working relationships with project counterparts.
• Understanding of the sub-national governance context and the operational realities of public administration entities is an asset.

 

Languages

• Excellent knowledge (spoken and written) of English is required.
• Knowledge of Dutch is a strong asset given the Flemish context.

 

How to Apply

Interested experts are invited to submit both a technical and a financial proposal (maximum 2 pages in total), in line with the requirements outlined in these Terms of Reference. The technical proposal should include at least two references to published or publicly available work demonstrating the applicant's ability to conduct legal or policy analysis of AI impact assessment tools, and to communicate complex findings accessibly for a practitioner or policy audience.

Proposals should be sent to s.jameson@unesco.org with the subject heading "Application for Applied Impact Assessment Guidance." 

The deadline for submission is May 8th  at 23:59 (Paris time).

Please note that no additional attachments are available, and any submissions which do not adhere to the above guidelines will not be reviewed.