Amsterdam builds trust in AI and gets feedback through algorithmic transparency

June 6, 2024

Amsterdam builds trust in AI and gets feedback through algorithmic transparency

In our increasingly algorithm-driven world, transparency and accountability have taken center stage. Amsterdam has been a pioneer in building human-centered AI and has used it for a long time to improve the accessibility and experience of the city’s services.

"Algorithmic transparency is essential to create trust among citizens and to receive valuable feedback", says Hanna Tak from the City of Amsterdam.

One notable response to this need is the Amsterdam Algorithm Register, a pioneering initiative developed in partnership with Saidot and the City of Amsterdam. Here’s how it was done.

The challenge: How to ensure AI is at the service of the people of Amsterdam – not the other way around

Taking a value-driven approach, Amsterdam started the work on building responsible AI very early on to ensure transparency and clarity in its use of algorithms. After all, algorithms used in public services must adhere to the same rules and principles as other public services.

That means they must treat people equally, not limit their freedom, and be transparent and open to democratic control.

As algorithms permeate various aspects of daily life, algorithms should never have the final say and should not work solely based on correlations. Therefore, it is crucial to document the underlying principles, decisions, and development processes behind AI.

The solution: An AI Register that’s one of the first of its kind in the world.

To give the people of Amsterdam access to understandable and current information on how algorithms affect their lives, the City launched a register of its algorithms as one of the first two cities in the world – alongside the City of Helsinki.

Saidot platform provided a standardised, archivable, and easily searchable method to document every aspect of algorithmic processes. This solution extended transparency not only to citizens but also to independent auditors and the public at large.

A collaborative effort improving the lives of many

The Amsterdam Algorithm Register was a collaborative endeavor envisioned and developed in partnership with the municipal government, aligning seamlessly with the city's commitment to responsible data and AI practices. It involved teams across different parts of the organisation: Development, communications, legal team and others.

The outcome

Since its launch in 2020, the Amsterdam Algorithm Register is an example of the City’s efforts to increase trust with its citizens and open AI up for feedback. It offers citizens a meaningful way to engage in the democratic process.

The City of Amsterdam uses algorithms for many use cases. For example, they can show if there’s a risk of a major traffic jam, or if an area with excessive danger that requires action.

The Register has inspired other organisations to follow Amsterdam footsteps with the initiatives of their own.

“AI registers can be a valuable tool for both organizations and the public to assess the legitimacy of AI solutions impacting people's lives,” says Hanna Tak from the City of Amsterdam.

AI continues to grow as an inseparable part of our everyday lives. So, what’s next?

Lately Amsterdam has been drafting their policy on generative AI. How can you strike a balance between experimenting with GenAI and maintaining control over its use?

Amsterdam will share learnings in the World AI Summit in Amsterdam on October 11^th, while hosting a workshop on this topic together with Saidot.

Want to learn more about algorithmic transparency?

For a deeper dive into how Amsterdam achieves algorithmic transparency using Saidot's platform, explore the Amsterdam Algorithm Register microsite.

June 6, 2024

Amsterdam builds trust in AI and gets feedback through algorithmic transparency

More Insights

Quick guide: How to identify, assess, and manage AI risks throughout the AI lifecycle

AI Act seen as a positive move by European businesses; concerns remain around how to comply

Saidot Launches Library of AI Resources to Help Businesses Implement AI Safely

Get started with responsible AI.