Artificial Intelligence and Machine Learning
This guideline supports CAs in integrating Artificial Intelligence (AI) and Machine Learning (ML) technologies into their procurement needs and processes. Considering that AI/ML solutions can enhance a wide range of applications, this document has the scope of providing a comprehensive, step-by-step approach to help CAs tailor existing procurement templates and incorporate AI/ML functionalities into solutions they consider appropriate.

1. Definition and applications of AI/ML tools to LDTs
Artificial Intelligence (AI) and Machine Learning (ML) can help cities in their decision-making process with advanced tools and data-driven models; artificial intelligence and machine learning technologies enable communities to analyse vast amounts of data, identify patterns, and derive actionable insights, thereby optimising urban management processes and fostering innovation. By integrating AI/ML into Local Digital Twins (LDTs), cities can create dynamic, responsive systems that adapt to changing conditions and support real-time decision-making.
AI/ML technologies have wide-ranging applications across smart city initiatives, including predictive maintenance, transportation optimisation and smart traffic management, air quality control and predictive modelling, resource management, and citizen engagement. AI-powered systems enhance operational efficiency by automating routine tasks and providing intelligent recommendations, while ML models continuously learn and improve their performance based on new data.
Based on the EU Artificial Intelligence Act, an “AI system means a machine-based system that is designed to operate with varying levels of autonomy and that may exhibit adaptiveness after deployment, and that, for explicit or implicit objectives, infers, from the input it receives, how to generate outputs such as predictions, content, recommendations, or decisions that can influence physical or virtual environments.”55 In other words, AI involves the development of systems capable of performing tasks that typically require human intelligence, such as reasoning, learning, and decision-making. AI offers transformative potential for cities, enabling automated processes and intelligent decision-making in areas such as traffic management, energy distribution, and public safety. AI systems can analyse diverse data sources, including IoT sensors and historical records, to provide communities with predictive insights and automated solutions tailored to their unique needs.
Machine learning, a subset of AI, focuses on developing algorithms that allow systems to learn from data and improve their performance over time. ML tools enable cities to extract valuable patterns and trends from complex datasets, facilitating data-driven decision-making across various domains. For example, ML models can predict energy consumption patterns, identify anomalies in infrastructure operations, and optimise public services based on historical and real-time data. By leveraging ML, cities can enhance the adaptability and efficiency of their digital ecosystems.
2. Understanding and identifying AI/ML-driven objectives and tools to procure
To successfully integrate AI and ML into LDT platforms or related solutions, it is essential for cities to start by clearly understanding their AI/ML-driven objectives. Identifying these objectives involves determining the specific needs and goals of the community, such as enhancing data analysis, improving decision-making, or automating workflows. These objectives will guide the selection of appropriate AI/ML tools and technologies, ensuring they align with the city's strategic vision.
Key steps to understanding and identifying AI/ML-driven objectives include:
- Assessment of current capabilities: Evaluate existing systems and processes to identify gaps where AI/ML can add value. For instance, determining whether advanced analytics, predictive modelling, or natural language processing capabilities are required.
- Defining use cases: Establish specific use cases where AI/ML can address operational challenges or improve service delivery. Use cases include traffic optimisation, energy efficiency, or citizen engagement through chatbots.
- Stakeholders’ involvement: Collaborate with relevant stakeholders, such as data scientists, IT teams, and decision-makers, to define priorities and ensure alignment with the city's needs.
- Technology evaluation: Identify the AI/ML tools, platforms, and algorithms most suited to the defined use cases. This includes evaluating factors like scalability, interoperability, and ease of integration with existing systems.
- Setting measurable goals: Define Key Performance Indicators (KPIs) to measure the success of AI/ML implementation. Examples of KPIs include accuracy improvements, cost savings, or user satisfaction metrics.
By following these steps, cities can effectively define their AI/ML-related objectives and procure tools tailored to meet their requirements.
3. Additional AI/ML requirements to be integrated into relevant procurement templates
Integrating AI/ML components into application services demands that cities consider additional requirements. These requirements ensure that AI/ML technologies are effectively embedded within solutions and deliver the intended benefits. The adaptation of templates for such procurements involves aligning the Service-Level Agreements (SLAs), technical background, functional, technical and legal requirements, and other relevant sections in the tendering documents to correctly include the unique capabilities of AI/ML.
Cities must carefully tailor the templates to address these aspects, ensuring that AI/ML solutions meet current requirements and remain flexible for future enhancements. By embedding these considerations into procurement templates, cities can drive the successful adoption of AI/ML technologies across their digital infrastructures. However, the following technical and functional requirements are intended to serve as a baseline reference for the CAs. These requirements are not exhaustive and may not fully address the specific needs or circumstances of every procurement context. CAs are advised to carefully review, adapt, and customise the provided requirements to ensure they align with the objectives, constraints, and operational priorities of their specific projects.
The following is an overview of the technical and functional requirements followed by legal frameworks.
Technical framework3.1 Technical & Functional requirementsTest requirements of AI/ML models and algorithms within LDT: Purpose:
Data:
Self-learning: AI/ML models and algorithms should be capable of learning and adapting based on real-time data inputs in order to support continuous improvement of the models and algorithms. Human oversight: AI system should incorporate solutions that enable human oversight to maintain human autonomy and reduce potentially negative impacts (e.g., via a combination of Human-in-the-loop (HIL), Human-on-the-loop (HOL), Human-in-command (HIC), or other approaches recognised and accepted as benchmarks) Explainability:
Traceability:
Transparency:
Fairness:
Robustness:
Performance:
Please consider that these requirements can be integrated within any template, except for network infrastructure and IoT devices templates, considering the necessary adaptations. 4.2 Minimum Interoperability Mechanisms (MIMs)While embedding procurement templates to guarantee the successful adoption of AI/ML technologies across digital infrastructures, cities should also consider the Minimum Interoperability Mechanisms (MIMs) to enable sufficient interoperability for data, systems, and services, specifically in smart city solutions. More specifically, cities can refer to MIM5: Fair and Transparent Artificial Intelligence 3 produced by OASC. At present, due to the 2024/25 MIMs redesigning framework, there is an ongoing transition towards a more comprehensive approach, which will focus on "Interoperable AI". For more specific information related to this topic, you can consult Guideline on Compliance with MIMs, standards, specifications and certifications. 4.3 Evaluation and Award criteriaThis section of the Guideline aims to provide insights and possible tender development options, addressing any gaps in specific technical and functional requirements for AI & ML solutions. Hence, for the evaluation and award criteria identification, selection and development, please refer to the technical and functional requirements mentioned above in section 3.1., as the minimum technical and functional requirements, de facto serving as Technical Specifications, to be then included in the chosen template among the available ones. Alternatively, if the most appropriate template is chosen from the available options, specific technical/functional requirements may not need to be provided. In such cases, additional discretionary criteria can be included in the tender evaluation & award criteria section (that is, for example: “Does the offer provide any AI & ML solutions for the given application as an add-on” – if a given tenderer provides AI & ML solutions, it will be assigned an additional rewarding score based on the proposal’s evaluation carried out by the CA itself). |
Legal frameworkWhile embedding procurement templates to ensure the successful adoption of AI/ML technologies, cities must also comply with the EU legal framework. This involves adhering to relevant EU laws, regulations, and guidelines to safeguard the ethical, transparent, and lawful integration of AI/ML in digital infrastructures. Of further assistance to cities and communities could be the consulting of Guideline on the EU Legal Framework on digital products and services relevant to LDTs, which offers an overview of compliance requirements for public procurement processes involving AI/ML technologies. Additionally, cities can refer to the EU model contractual AI clauses, available on Living-in.EU, to ensure standardisation and legal alignment when drafting contracts for AI solutions. These resources provide practical tools for navigating the legal landscape and fostering trust and transparency in AI deployment. |
- 1Rudin, C., & Radin, J. (2019). Why Are We Using Black Box Models in AI When We Don’t Need To? A Lesson From an Explainable AI Competition. Harvard Data Science Review, 1(2). https://doi.org/10.1162/99608f92.5a8a3a3d
- 2High-Level Expert Group on Artificial Intelligence. (2019). ETHICS GUIDELINES FOR TRUSTWORTHY AI. European Commission: https://op.europa.eu/en/publication-detail/-/publication/d3988569-0434-11ea-8c1f-01aa75ed71a1
- 3https://mims.oascities.org/mims/oasc-mim5-transparency