Does your city use artificial intelligence? If so, how? The growing role and use of algorithmic transparency registries
Nine European cities have developed a free-to-use, open-source “data schema” for city algorithm registries. These cities include Barcelona, Bologna, the Brussels-Capital Region, Eindhoven, Mannheim, Rotterdam and Sofia based on the example of Amsterdam and Helsinki.
Transparency in artificial intelligence systems is a hot topic. The UK Data Ethics and Innovation Center has updated the information An algorithmic transparency storage standardbased on their report of bias algorithmic decision making. We can expect transparency to be discussed as part of this inquiry In the UK on what good AI governance looks like. We can also expect an EU database for risky AI systems as proposed by the EU EU AI law (Article 60).
Here we briefly explain the issue of transparency in the use of artificial intelligence systems by public bodies and point out the experiences of cities developing algorithmic transparency registers.
How do cities use artificial intelligence?
Examples in Amsterdam are reporting problems (e.g. garbage or maintenance problems), illegal vacation rentals and parking enforcement. You can watch them Register here.
However, it is difficult to determine when and where authorities are using AI, as the use of AI systems is not always publicized.
Lack of transparency is a problem. It is one of the limitations mentioned in studies on the use of AI by public authorities (for example, the report on canceled AI projects in the public sector here). Without transparency, citizens do not know when and where their data is used, how it is used, and how to seek corrections if problems arise.
The growing role of transparency registers
Some authorities have developed AI/algorithmic decision-making transparency registries to address the current lack of transparency about when and where AI is used.
The report, which explains how nine European cities have developed them, explains the AI registry as follows:online list of these different algorithms with a brief explanation of where, how and why they are used.“
Developing such a registry is not simple. It requires consideration of questions including: What information is needed when we want to collect data? How can we collect meaningful, comparable information that increases the efficiency of electronic administration instead of difficulties? And how do we use and manage data in different applications?
Nine cities have developed a “data schema” now that the use of artificial intelligence is in its infancy. It is better to agree on common standards now, before more advanced artificial intelligence is introduced more widely in public services.
The goal is that the common approach is shared and copied by other cities that otherwise might not have the resources to invest in their own initiatives.
What problems arise in transparency registers?
The experiences of cities developing or adopting transparency registries are useful in highlighting topics that are important in many scenarios when implementing AI – in public or private organizations. For example:
- which systems should and should not be included in the register: Pasi Rautio, service manager of Helsinki’s strategy department, said “Almost 40,000 people work in the city… and we have a lot of projects going on. So we don’t know if we have fully registered all our algorithms, and there is also the question of the definition of artificial intelligence and what exactly should be included in the register.
- need expert groups in the organization. The report states: “One way Helsinki has contributed to a better approach to this is by creating specific teams in the city administration that work on targeted aspects of digital transformation, such as artificial intelligence or data and analytics. That way, when other colleagues in the city administration start working on something that might involve artificial intelligence, they are now much more likely to do so together with a team of experts and therefore know that any algorithm is shared in the registry.“
- the need for an artificial intelligence advisory board. Rotterdam’s Innovation Officer explained that “we created an external Algorithm Advisory Board to review what we’re doing and whether we’re doing it right. The idea is to remain transparent to all our stakeholders, regardless of whether they are in the administration or external.“
- questioning what information the public needs for transparency. The Brussels Smart City and Innovation Officer noted the following challenges: “Finding the right balance of providing enough information about the algorithm without overburdening people, and ensuring that their use is not viewed as an Additional Burden but as a public benefit, are all on the minds of administration officials.“
complex algorithms in automated systems trained on biased data can transfer bias to groups of citizens. Algorithm registries provide transparency about the development and implementation of algorithms and provide an important safeguard for the responsible use of artificial intelligence.