Digital courage & the surveillance of garbage
Smart cameras on garbage trucks to assess the garbage in the city. Artificial intelligence often seems like a solution. But information about garbage is not the problem that requires a solution, the problem is that people throw their garbage on the street. Meanwhile, it takes an enormous amount of energy to train algorithms, private companies are entering our lives more and more deeply and 'all-seeing' cameras in public space are by no means neutral. Reasons for the Digital Courage and Fearless Tech Expo.
The digital infrastructure in our cities is are more and more being outsourced to tech companies. As a result, local governments and public services are increasingly dependent on software as a service, allowing the companies that make this software to enter more and more the spheres of our lives. The risks and threats we face as part of this cycle are not being addressed by local politics. That is why we want to continue to research the impact of new technology on cities, from a political-economic point of view. And we are building the Digital Courage and Fearless Tech Expo.
A full Boeing
A good example of how new technologies are presented as a solution, but in fact cause new problems, is artificial intelligence (AI) and machine learning. Take BERT, an algorithm that teaches computers how to use and recognize language. Training this algorithm requires about the same amount of energy as a return flight of a full Boeing 747 between New York and San Francisco.
Amsterdam teaches AI to recognize certain types of gabrage on the street, such as bin bags and cardboard boxes. The idea is to install smart cameras on garbage trucks so the AI can learn to assess the garbage on the street, giving the government an increasingly accurate overview of how messy the city is.
This project costs a lot of resources, because that AI needs to be trained, but the problem that this offers a solution for seems non-existent: we already know where the garbage is, we can see it with our own eyes. The problem is not information, the problem is people throwing their garbage on the street. If information is not the problem, data collection can not be the solution.
Moreover, there is cost to this solution: teaching recognition cameras to monitor public space, whether or not in search of waste, can also be given other functions, so-called function creep. The all-seeing camera computers can be used for surveillance of other situations or people. Is it worth turning the public space into a panopticon, monitored by machine learning cameras, to get rid of trash bags lying around?
Watch part 1 of Digital Courage Online Expo: Uber Eats
And find more events on digital democracy by Cities for Change
Discover the online expo (in Dutch) Fearless Tech & Digitale Moed