AI technologies offer significant potential to improve the effectiveness and efficiency of public administrations and are key to achieving our Digital Decade ambitions. Ranging from chatbots to AI-powered local digital twins offering complex modelling capabilities for city planners, AI and its potential uses are developing all the time.
Data governance, access to high volumes of quality data, and technical interoperability between platforms and systems are key areas of importance being addressed through Living-in.EU and other smart communities initiatives. However legal, procurement and transparency tools are equally important as ways to foster update of AI by public authorities in a way that ensures a high degree of trust among citizens. Cities and communities in the EU are leading the way and offer use cases that can be scaled up through cooperation.
This workshop will outline three strands of work (see below) that are getting underway through Living-in.EU and its partner organisations Eurocities and OASC, as well as broader developments supported by the European Commission in the context of public sector AI, including new projects funded through the DIGITAL Programme. This workshops aims to presenting these with a view to aligning them for maximum impact and complementarity. It is also an open call to interested regions, cities and communities to contribute to this work. Please join to share your experiences and ideas for collaboration!
To attend, please register here. The meeting link will be issued to registered participants.
14.00 Welcome by moderator, Martin Brynskov, OASC (Open and Agile Smart Cities)
14.05 EU Policy Context including the AI Act and the Coordinated Plan on AI, Yordanka Ivanova, DG CONNECT UNIT A2, Artificial Intelligence
14.20 Initiatives Carousel – series of short presentations on initiatives underway, followed by a discussion on alignment and next steps
- Contract clauses for public procurement of AI (Amsterdam, DG GROW, Living-in.EU Legal & Ethics sub-group), Lydia-José Prinsen, Office of the CTO, City of Amsterdam and Ivo Locatelli, Senior Expert, DG GROW, Unit C2 Procurement
- Algorithm registries for transparent AI (Eurocities Digital Forum Lab), Lodewijk Noordzij, Policy Advisor, Living-in.EU Legal & Ethics Sub-group
- MIM5 for Fair AI, Michael Mulquin, OASC, Living-in.eu Technical subgroup
- AI Testing and Experimentation Facilities for smart cities and communities, Evangelia Markidou, DG CONNECT, Unit A1, Robotics and AI Innovation and Excellence
- Living Labs for AI Regulatory Learning (JRC and ENoLL), Sven Schade, JRC, Digital Economy Unit B6 and Fernando Vilariño Freire, Associate Director at Computer Vision Center, Barcelona
- Open Data for AI, Martino Maggio, Engineering
- CommuniCity- AI responding to the needs of Cities and Communities, Katarina Makrogamvraki, University of Amsterdam
15.05 Discussion, including interdependencies between work, open floor format, led by moderator
15.50 Conclusions and next steps
Model Contractual Clauses for the Procurement of AI
Amsterdam developed a set of contractual clauses for the procurement of Artificial Intelligence, to create a framework for the information that suppliers need to provide on algorithms used. The clauses build on the Ethics Guidelines developed by the European Commission’s (EC) High Level Expert Group on Artificial Intelligence (AI), and the city’s own strong digital ethics policy and work though the Cities Digital Rights Coalition. The contractual clauses are now being adapted as a template for other public administrations to reuse. DG GROW has held two webinars to date on the clauses. The first (June 2021) showcased Amsterdam’s approach to AI procurement and the second (May 2022) presented the adapted clauses, partially aligned with the AI Act currently under negotiation. The draft clauses are now available on Living-in.EU for feedback. See here for more information.
From early 2023, DG GROW’s Digital Public Buyers Initiative and Living-in.EU will work with stakeholders to trial the adapted Amsterdam procurement clauses and to develop guidance material. When finalised, these clauses will be a useful tool for public procurers at local, regional and national level to procure ethical, human-centric AI, to boost uptake of AI and to increase citizen trust in its use.
AI Algorithm Registries
An AI Algorithm Registry is an online tool or website that provides citizens with information on artificial intelligence systems used by public administrations in reaching decisions affecting them in a transparent and understandable manner. The Cities of Amsterdam and Helsinki are examples of existing Algorithm Registries.
Eurocities’ Digital Forum Lab is working with a group of cities to develop and pilot a common data schema for cities that wish to create their own interoperable local AI Registries, drawing on the Amsterdam model contractual clauses, the Dutch data model for algorithm registries, the UK Algorithmic Transparency Standard and others. The data schema will not only serve cities that would use an external supplier to develop a solution for them, but could also help those that would want to do it in-house or through open-source software. See here for more information.
Fair AI Minimum Interoperability Mechanism (MIM 5)
OASC, which leads the Living-in.EU Technical Sub-group, is beginning to work with cities to develop a Fair AI MIM, also known as MIM 5 (Minimal Interoperability Mechanism), part of the MIMs Plus set of technical specifications. This technical tool will assist cities to automatically (i) check that AI service providers have supplied the appropriate information and safeguards in line with the transparency requirements, such as those set out in the standard contractual clauses, and (ii) link this data to an AI registry on how AI was used in decisions affecting citizens.