Photo by @alexmotoc

Algorithm Registries in Amsterdam and Helsinki

New algorithmic governance in cities within Europe shows promise in 2020

There are growing concerns with the effect algorithmic governance has on society. Recent hit documentaries like The Social Dilemma is one aspect, however there is far more to consider. America (mainly Trump) banning TikTok and WeChat from America. Ethical concerns of artificial intelligence are becoming a growing discussion. AI for children has been explored by UNICEF. Nations across the world are making national and multilateral AI strategies, many include discussions ethics or consequences of the applications of technology.

Discussions on these levels are great, of course, and yet there is a very practical dimension to algorithmic governance.

Still, what do they mean and what practical implications do they have?

For example in cities.

That is perhaps why what Amsterdam and Helsinki is doing is so interesting.

Khari Johnson writes in VentureBeat published the 28th of September about the launch.

Amsterdam and Helsinki has launched algorithm registries to bring transparency to public deployments of AI.

City AI registry: to detail how each city government uses algorithms to deliver services.

Amsterdam and Helsinki launched AI registries to detail how each city government uses algorithms to deliver services.

They can both be found here:

These registries were introduced as part of the Next Generation Internet Policy Summit.

This is organised in part by the European Commission and the city of Amsterdam.

To be fair, it does not include a great deal of algorithms yet.

However, it might give an idea or indication of what transparency and responsibility in the context of a city governing with algorithms could look like.

In this manner citizens or people in companies could themselves to some extent audit or challenge the algorithms (decision-making) or consequences.

According to Johnson datasets used to train a model and a description of how an algorithm is used is included.

How do humans utilise predictions, and how are algorithms assessed for potential bias or risks?

Perhaps by local governments!

At least this is a fascinating idea of how more responsible algorithmic governance could look like.

Johnson writes:

“A complete algorithmic registry can empower citizens and give them a way to evaluate, examine, or question governments’ applications of AI.”

Afterwards she mentions that U.S has created automated decision systems in 2017 to assess city use of algorithms and that a report last year found a lack of transparency and inability to access information about algorithms used by city government agencies.

Poor performance or bias can lead to distrust between government and citizens. According to Johnson this was cautioned by a joint Stanford-NYU study analysing how the U.S. government deploys AI systems.

Governments and business leaders are increasingly applying advanced algorithms to automate tasks or make decisions.

Johnsons refers to examples in the Netherlands and the United Kingdom where use of algorithms

“A Dutch court ordered government officials to stop using the SyRI algorithm due to discrimination against immigrants and people living in low-income households. As was the case with facial recognition in the U.K. earlier this year, a judge deemed use of the algorithm a violation of human rights.”

Although we see algorithmic governance as an issue for cities, it is increasingly also of national and multilateral interest. Johnsons mentions four different initiatives:

  1. “Finland joined 12 other countries to form the AI Partnership for Defense. Hosted by the U.S. Department of Defense’s Joint AI Center, the coalition will discuss how to translate ethics or policy principles into practice.
  2. The same week, EU Commission members began talks with high-level Chinese officials about issues important to the global economy including AI.
  3. The European Laboratory for Learning and Intelligent Systems (ELLIS) launched the second wave of its pan-European initiative to accelerate AI research in dozens of cities across the EU, continuing the $220 million initiative to keep AI talent in Europe.
  4. In another example of Finland attempting to create a more democratic approach to AI, last December government officials made AI training available with the stated goal of training at least 1% of EU citizens about the fundamentals of AI.”

I currently find myself in Norway.

Here a regulatory sandbox is being set up.

However, I wonder if it would be interesting as well to follow the good example from Helsinki and Amsterdam in Oslo.

Why not learn from neighbours?

This is #500daysofAI and you are reading article 485. I am writing one new article about or related to artificial intelligence every day for 500 days.

AI Policy and Ethics at www.nora.ai. Student at University of Copenhagen MSc in Social Data Science. All views are my own. twitter.com/AlexMoltzau