Pilot Project: Assessment for Responsible Artificial Intelligence together with Rijks ICT Gilde -Ministry of the Interior and Kingdom Relations (BZK)- and the province of Fryslân (The Netherlands)

Original Full text available from the Rijks ICT Gilde -Ministry of the Interior and Kingdom Relations (BZK)- web site in Dutch.

Artificial Intelligence (AI) is in more and more aspects of our lives. The technology – based on algorithms and data – is in numerous devices and can be useful in solving social issues. For example, about energy, sustainability or poverty. As a government, we have an exemplary role. If we want to seize the opportunities of AI, then important questions about ethics, technology, transparency and the possible effects of AI applications on our society must be answered.

The pilot “Assessment for Responsible AI” is a step in this process. During a three-month pilot, the Rijks ICT Gilde (Ministry of the Interior and Kingdom Relations (BZK)) , in cooperation with the province of Fryslân and AI authority Prof. Dr. Zicari and his team, is investigating a deep learning algorithm in practice.

Reason for the pilot
During the conference “AI and the future of Europe” in Brussels on March 30, 2022, Secretary of State Alexandra van Huffelen told us that the digital transition and the use of AI should always be human-centered and based on our democratic values and rights. Governments should lead by example in this regard.

As a government, we want to seize the opportunities of AI, but the technology still raises many important questions. How reliable are algorithms? Can an algorithm discriminate? What are the ethical and social effects of AI and how transparent is its use? In addition, the use of AI must always be human-centered and based on our democratic values and rights.

With that comes an impressive amount of rules, frameworks and frameworks in the field of AI. How do you apply them in practice? What do you need to pay attention to? And how do you integrate them into the development and use of AI?

With the pilot “Assessment for responsible AI” we hope to get answers to these questions and more. First, to stimulate awareness and dialogue about AI within government. And then to be able to confidently deploy the technology for the questions of tomorrow.

Background for the pilot

During this three-month pilot, the practical application of a deep learning algorithm from the province of Fryslân will be investigated and assessed. The algorithm maps heathland grassland by means of satellite images for monitoring nature reserves. The testing of this algorithm is done in collaboration with an international interdisciplinary team, based on the ‘Z-Inspection® method’ – a process to test AI for reliability.

This involves testing the algorithm for social, ethical, technical and legal implications. This is done in interdisciplinary teams according to a holistic methodology: it looks at the coherence, arrangement and interaction of the features to a system. The holistic nature of the method leaves room for different dimensions and views, and focuses in particular on the interpretation and discussion of ethical issues and tensions.

Pilot objectives
The pilot has multiple objectives, so each party can get its own win from the pilot:

A science-based assessment in practice based on a concrete AI application;
To understand how to carefully and responsibly organize your processes for developing and using AI;
Learning which frameworks, laws, and regulations are important and must be tested in the different phases of development and use;
Increasing insight and overview through recommendations;
Stimulating dialogue and increasing awareness about applying reliable AI.

Z-Inspection® : A method for responsible AI
One of the main characteristics of the Z-Inspection® method is its interdisciplinary nature. The complexity of an AI application is reflected in the composition of the team. The diversity of participants provides a more inclusive assessment of the reliability of an AI application.

Another strong feature of this method is its dynamic application. A standard checklist does not adapt to the case. In contrast, the holistic approach determines which issues are central to the use case (real-world application) at different stages of the process. And assesses which aspects of the case are most important, moving back and forth between intra- and interdisciplinary discussions.

With the application of this method, awareness grows within the own organization, a frame of mind (mindset) emerges for responsible data use and the reliable application of AI. Both Province of Fryslan and the Rijks ICT Gilde are testing the method in practice to learn together how to look at an AI application from different dimensions.

Pilot outcome
Using the Z-Inspection® method, the advantages and disadvantages of the AI application under investigation are described. This is done using an ongoing and iterative (repeating and increasingly refining) research process. Participants are hereby given space to openly reflect and document what is known (and unknown) about the capabilities of an AI application as a basis for later evaluations.

Participating parties
Several parties are working together in the pilot. The Province of Fryslân, AI authority Prof. Dr. Zicari and his team and the UBR| Rijks ICT Gilde are jointly investigating the reliability of AI applications and their responsible use. Leeuwarden municipality, the University of Groningen/Campus Fryslân and policy advisors of the Ministry of the Interior and Kingdom Relations are participating as observers.

Read more about the participating parties below.

Province of Friesland
During the pilot project, a deep learning algorithm of the Province of Fryslân will be evaluated. This algorithm visualizes heather grassing for monitoring nature reserves. The Province of Fryslân is investing in the coming years in the smart and effective use of data. The province sees that almost all provincial developments and social tasks contain a data component. This creates urgency in the subject. To be able to responsibly respond to technological developments as a province, a sharp vision on data and AI is needed. Participation in the pilot helps design the future digital infrastructure and outline ethical frameworks.

Read more > https://www.fryslan.frl/

UBR | Rijks ICT Gilde
The Rijks ICT Gilde (RIG) (Ministry of the Interior and Kingdom Relations (BZK) ) is an ambitious tech organization that implements projects across the central government. The organization uses its knowledge and network to create smart partnerships and solutions. It employs specialists with a drive to help the Netherlands move forward in the fields of data, software and security.

Mycelia is a young, dynamic and energetic programme of the Rijks ICT Gilde that uses the responsible use of data and AI to achieve impact. We do this by answering and filling in relevant questions without compromising the public values and fundamental rights that the government stands for.

With enthusiasm and specific knowledge, Mycelia’s data and AI experts work on impactful and ethically responsible projects for the public sector and collaborate with organizations and experts in order to learn, grow and develop from each other.

With the growth of data and the rise of AI, we see a government that will never be the same again. Responsible use of AI and trust is and will therefore become increasingly important. We believe in helping and supporting each other and an honest and transparent government. We want to leave a better world for the next generations.

In the pilot, the RIG provides expertise on AI, ethics and responsible data use within government.

Read more > https://www.ubrijk.nl/service/rijks-ict-gilde

Translated with www.DeepL.com/Translator (free version)

Update (January 25, 2023)

Timeline
The Assessment for Responsible AI pilot took place from May 2022 through January 2023.

Project Members:

Sara M. Beery,
Marjolein Boonstra,
Frédérick Bruneault,
Subrata Chakraborty,
Tjitske Faber,
Alessio Gallucci,
Eleanore Hickman,
Gerard Kema,
Heejin Kim,
Jaap Kooiker,
Ruth Koole,
Elisabeth Hildt,
Annegret Lamadé,

Emilie Wiinblad Mathez 
Florian Möslein,
Genien Pathuis,
Rosa Maria Roman-Cuesta,
Marijke Steege,
Alice Stocco,
Willy Tadema,
Jarno Tuimala,
Isabel van Vledder,
Dennis Vetter,
Jana Vetter,
Elise Wendt,
Magnus Westerlund,
Roberto V. Zicari

Announcing a new Pilot project with the Province of Friesland and UBR Rijks ICT Gilde (part of the Ministry of the Interior and Kingdom) the Netherlands.

Excited to announce that on May 16, 2022 we had our kick-off of the pilot project “Assessment for responsible AI” together with the Province of Friesland (Fryslân), a team of the  Z-inspection® initiative, and the UBR  Rijks ICT Gilde (part of the Ministry of the Interior and Kingdom Relations, the Netherlands).

Together we will investigate the reliability of AI applications for the Province of Friesland and their responsible use, using the  Z-inspection® process and the EU Framework for Trustworthy AI.

The Leeuwarden Municipality, University of Groningen/Campus Fryslân and policy advisors from the Ministry of the Interior were also invited as observers.

Announcement in Dutch :

Kick-off krachtig samenwerkingsverband voor verantwoorde AI

Nieuwsbericht | 17-05-2022 | 13:42- Rijks ICT Gilde. 

English Translation:

Kick-off powerful partnership for responsible AI

News release | 17-05-2022 | 13:42
Yesterday was the kick-off of the pilot ‘Assessment for responsible AI’. The Province of Fryslân, AI authority Prof. Dr. Zicari and the UBR Rijks ICT Gilde are jointly investigating the reliability of AI applications and their responsible use. Leeuwarden municipality, Groningen University/Campus Fryslân and policy advisors from the Ministry of the Interior are invited as observers.

Artificial Intelligence (AI) is appearing in more and more aspects of our lives. It is in all kinds of devices we use in our work and private lives. The technology, based on data and algorithms, can be useful in solving social issues about energy, sustainability or even poverty, for example.

As a government, we want to exploit the opportunities of AI, but the technology still raises many important questions.

How reliable are algorithms? Can an algorithm discriminate? And how transparent is the use of AI?

In the three-month pilot ‘Assessment for responsible AI’ we are looking for answers to the questions:

How do you as a government steer the development and use of responsible AI?
What frameworks, laws and regulations are important, and how do we test them in the development and use of AI?
How do you analyze, assess and improve AI applications? And are the applications in line with public values and human rights?

In the pilot we assess an algorithm of the province of Fryslân. We will analyze it using the Z-inspection® method of Roberto Zicari; a self-assessment in which participants discuss critical issues such as: the purpose of the algorithm, the development process, ethical dilemmas and conflicts of interest. The Z-inspection® method is a working method to analyze, assess and improve AI applications in a sustainable, demonstrable and transparent way. This enables organizations to develop and use responsible AI applications in a structured way.

Furthermore, it is very important that the knowledge and experiences from the pilot are shared. First of all, to stimulate digital awareness and dialogue about AI within the government. And then to be able to confidently deploy the technology for the questions of tomorrow.

Lessons Learned: Co-design of Trustworthy AI. Best Practice. By Helga Brogger, President of the Norwegian Society of Radiology

Mission: …Aid the development of designs with reduced end-user vulnerability…

-“…Socio-technical scenarios can be used to broaden stakeholders’ understanding of one’s own role in the technology, as well as awareness of stakeholders’ interdependence…”

– “…Recurrent, open-minded, and interdisciplinary discussions involving different perspectives of the broad problem definition….”

– “…The early involvement of an interdisciplinary panel of experts broadened the horizon of AI designers which are usually focused on the problem definition from a data and application perspective…”

– “…Consider the aim of the future AI system as a claim that needs to be validated before the AI system is deployed..”

-“…Involve patients at every stage of the design process … it is particularly important to ensure that the views, needs, and preferences of vulnerable and disadvantaged patient groups are taken into account to avoid exacerbating existing inequalities…”

Thank you, Roberto V. Zicari and the rest of the team for these insights!

— Helga Brogger, President of the Norwegian Society of Radiology

………………………………………………………………………………………………………………………………………………………………………….

Co-Design of a Trustworthy AI System in Healthcare: Deep Learning Based Skin Lesion Classifier.

Roberto V. Zicari, Sheraz Ahmed, Julia Amann, Stephan Alexander Braun, John Brodersen, Frédérick Bruneault, James Brusseau, Erik Campano, Megan Coffee, Andreas Dengel, Boris Düdder, Alessio Gallucci, Thomas Krendl Gilbert, Philippe Gottfrois, Emmanuel Goffi, Christoffer Bjerre Haase, Thilo Hagendorff, Eleanore Hickman, Elisabeth Hildt, Sune Holm, Pedro Kringen, Ulrich Kühne, Adriano Lucieri, Vince I. Madai, Pedro A. Moreno-Sánchez, Oriana Medlicott, Matiss Ozols, Eberhard Schnebel, Andy Spezzatti, Jesmin Jahan Tithi, Steven Umbrello, Dennis Vetter, Holger Volland, Magnus Westerlund and Renee Wurth.

Front. Hum. Dyn. |Human and Artificial Collaboration for Medical Best Practices, July 13, 2021

VIEW ORIGINAL RESEARCH article

Learn more

Our paper ” On Assessing Trustworthy AI in Healthcare Best Practice for Machine Learning as a Supportive Tool to Recognize Cardiac Arrest in Emergency Calls.” has been accepted for publication in Frontiers in Human Dynamics

The legislative proposal for AI by the European Commission has been published today.

The highly anticipated legislative proposal for AI by the European Commission has been published today.

Read the EU Regulatory Proposal on AI:

https://digital-strategy.ec.europa.eu/en/library/proposal-regulation-european-approach-artificial-intelligence

EU Press Release

Press release 21 April 2021 Brussels

Europe fit for the Digital Age: Commission proposes new rules and actions for excellence and trust in Artificial Intelligence

Kick off Meeting (April 15, 2021) Assessing Trustworthy AI. Best Practice: Deep Learning for predicting a multi-regional score conveying the degree of lung compromise in COVID-19 patients. In cooperation with Department of Information Engineering and Department of Medical and Surgical Specialties, Radiological Sciences, and Public Health – University of Brescia, Brescia, Italy

On April 15, 2021 we had a real great kick off meeting for this use case:

Assessing Trustworthy AI. Best Practice: Deep Learning for predicting a multi-regional score conveying the degree of lung compromise in COVID-19 patients.

71 experts from all over the world attended.

Worldwide, the saturation of healthcare facilities, due to the high contagiousness of Sars-Cov-2 virus and the significant rate of respiratory complications is indeed one among the most critical aspects of the ongoing COVID-19 pandemic
The team of Alberto Signoroni and colleagues implemented an end-to-end deep learning architecture, designed for predicting, on Chest X-rays images (CXR), a multi-regional score conveying the degree of lung compromise in COVID-19 patients.

We will work with Alberto Signoroni and his team and apply our Z-inspection® process to assess the ethical, technical and legal implications of using Deep Learning in this context.

For more information: https://z-inspection.org/best-practice-deep-learning-for-predicting-a-multi-regional-score-conveying-the-degree-of-lung-compromise-in-covid-19-patients/

This AI detects cardiac arrests during emergency calls

Jointly with the Emergency Medical Services Copenhagen, we completed the first part of our trustworthy AI assessment.
A ML sytem is currently used as a supportive tool to recognize cardiac arrest in 112 emergency calls.
A team of multidisciplinary experts used Z-Inspection® and
identified  ethical,technical and legal issues in using such AI system.
This confirms some of the ethical concern raised by Kay Firth-Butterfield, back in June 2018….

This is another example of the need to test and verify algorithms,says Kay Firth-Butterfieldhead of Artificial Intelligence and Machine Learning at the World Economic Forum.

“We all want to believe that AI will ‘wave its magic wand’ and help us do better and this sounds as if it is a way of getting AI to do something extremely valuable.
“But,” Firth-Butterfield added, “it still needs to meet the requirements of transparency and accountability and protection of patient privacy. As it is in the EU, it will be caught by GDPR, so it is probably not a problem.” However, the technology raises the fraught issue of accountability, as Firth-Butterfield explains. Who is liable if the machine gets it wrong? the AI manufacturer, the human being advised by it, the centre using it? This is a much debated question within AI which we need to solve urgently: when do we accept that if the AI is wrong it doesn’t matter because it is significantly better than humans. Does it need to be a 100% better than us or just a little better? At what point is the use, or not using this technology negligent?

Source: https://www.weforum.org/agenda/2018/06/this-ai-detects-cardiac-arrests-during-emergency-calls/

Image: CPR

The full report is submitted for publication. Contact me if you are interested to know more. RVZ

Resources:

Article World Economic Forum, 06 Jun 2018.

Download the Z-Inspection® Process

“Z-Inspection®: A Process to Assess Ethical AI”
Roberto V. Zicari, John Brodersen, James Brusseau, Boris Düdder, Timo Eichhorn, Todor Ivanov, Georgios Kararigas , Pedro Kringen, Melissa McCullough, Florian Möslein, Karsten Tolle, Jesmin Jahan Tithi, Naveed Mushtaq, Gemma Roig , Norman Stürtz, Irmhild van Halem, Magnus Westerlund.
IEEE Transactions on Technology and Society, 2021
Print ISSN: 2637-6415
Online ISSN: 2637-6415
Digital Object Identifier: 10.1109/TTS.2021.3066209
DOWNLOAD THE PAPER

The Z-Inspection® Process is available for Download!

The Z-Inspection® Process is available for Download! “Z-Inspection®: A Process to Assess Ethical AI”Roberto V. Zicari, John Brodersen, James Brusseau, Boris Düdder, Timo Eichhorn, Todor Ivanov, Georgios Kararigas , Pedro Kringen, Melissa McCullough, Florian Möslein, Karsten Tolle, Jesmin Jahan Tithi, Naveed Mushtaq, Gemma Roig , Norman Stürtz, Irmhild van Halem, Magnus Westerlund.IEEE Transactions on Technology and […]

Presentation on Z-Inspection® at the [AI4EU] Trustworthy AI workshop. November 13, 2020

Roberto V. Zicari did a 30 minutes presentation at the [AI4EU] Trustworthy AI workshop  on our research on Z-Inspection® , a process to assess Trustworthy AI.

YouTube: Link to when the presentation starts