Follow

Quayside, in Toronto, was supposed to be the leading proof of concept for smart cities in the Western world.

Google's parent company planned to invest almost $1B in autonomous garbage collection, self-driving taxis, and an extensive data collection layer - from bench occupation, to pedestrian crossing monitoring, to public transport live usage data.

The problem is that Alphabet failed to address all the concerns of the locals about privacy and data governance - and in many cases it even actively dismissed or derided them. The local government even mentioned several episodes of arrogance shown by Alphabet's representatives. So, two years and a half after its start, the project has been officially terminated.

This is a good example of how citizens can push back on surveillance capitalism. Smart cities have plenty of potential to improve lives. Collecting more accurate live data leads to better understanding of the problems and therefore better governance. But the collected data is extremely valuable, and it shouldn't solely lay in the hands of a private corporation. Especially if that corporation dismisses valid concerns about user privacy and data gatekeeping as technophobia.

These technologies should be controlled by elected officials, because they are the ones accountable (through the democratic process) for their correct usage. Representatives that misuse the data, or sell private citizens' data to 3rd-parties, are likely to be voted out. You can't say the same about Alphabet. Nobody voted for them, nobody can vote them out if they misuse the data, they aren't accountable to anyone other than their shareholders, and therefore they are not eligible for handling something as precious as the data flows of our cities.

And there's also a disturbing lesson to be learned here: companies like Google would rather kill billion-dollar projects and not improve our cities, rather than losing control over the data, or even just starting a conversation about data usage and accountability. Their interest is not to make the world a better place: it's to maximize profit. Improving things sometimes comes as an aftermath of their profit-seeking strategies, but sometimes it doesn't.

news.slashdot.org/story/22/08/

@blacklight capitalism surveillance is as frightening imho when its done by the gov or muni. china definitely comes to mind as an example, but even just the thought of so much data in the hands of average city cops is extremely distressing to me.

@blacklight also, there was a whistleblower in the usa that showed how the gov and the corps work together as one body with different limbs.

@nebuchi that's why accountability, transparency and data governance are all important for such projects to succeed in a sustainable way.

Accountability means that people elect representatives through the democratic process, the representatives have the power of collecting and handling the data, but if they misuse it (e.g. sell private citizens' data to 3rd-parties, or use it to enact discriminatory or persecutory practices) the citizens can still vote those representatives out. This is something that China doesn't have - and if nobody can vote you out for misusing their data, then you will misuse their data.

Transparency means that the same data is accessible both to the government and to scientific researchers - possibly in aggregate and anonimized form. That enforces both constructive two-way feedback loops that are mostly missing today, and it ensures that the decisions of the government are transparent and can be backed by data accessible to anyone in the scientific community.

Data governance means that the government and data experts must be involved in selecting what data needs to be stored - if the purpose of my sensors is to track volumes at traffic intersections, do I really need to store the pictures of the people on the database too? if it is to track usage of public transport against date and time, do I really need to store the personal ID of each NFC badge used to check-in? - as well as for how long the data can be stored, and how to transparently audit when and why it's accessed.

There's a lot of untapped potential in these areas, but we need to make sure that we clearly define the boundaries of what can be done. Both Toronto (surveillance capitalism) and China (surveillance totalitarianism) are examples of how these things should *not* be done.

Since 'AMGAF' have so much data about their customers, jurisdictions should impose taxes accordingly e.g. no "registration in Eire because of its lowest corporation tax rate"; instead, Canada has X million AMGAF customers, therefore those "big tech" companies must pay Y amount each year to operate in that country. Of course, EU and USA companies lobby to prevent such equity.

@citc this problem wouldn't exist if only those companies paid their taxes in the same country where they made their profits. That's something that several policymakers have tried to pursue without success for a few years. The argument is that it's hard to track all the steps in a digital value chain to tell where a company is supposed to pay their taxes. My argument is that it's not that hard, and these companies have already set up much more complex tracking processes - it's just that setting up one that would increase their tax burden is not in their interest.

@blacklight That's all cool and right and whatnot, but you lost me at "elected officials are accountable". In what world?

Elected officials protected the NSA, hunted down Edward Snowden, tortured Chelsea Manning, imprisoned Reality Winner, etc etc. Democratic shmemocratic, the whole "holding accountable" stuff has failed.

These technologies shouldn't be controlled by elected officials, these technologies should be banned, burned wherever we find them, and stigmatised so much that anyone even proposing them can never find work in tech for the rest of their lives.

@amberage @blacklight Elected officials are far from perfect, and there's issues with the system, but accountability to the public is, at minimum, a goal. Corporations have all of the same issues, except they're generally not even supposed to be accountable to the public. Lesser of two evils.

@ocdtrekkie @blacklight I'd prefer neither evil tbh. Why choose between bad and worse when you can just refuse to let either have it?

@amberage @blacklight Specifically on your examples, those were all individuals who chose to work for the intelligence community in the first place, and one is currently hiding in a country trying to obliterate Ukraine, after saying a bunch of stupid stuff on the topic.

@amberage @blacklight One problem with elected representative democracy is that the elected representatives are a single point of failure (through ignorance and/or corruption) with concentrated power over an extended portfolio with little or no expertise in all the fields that they're dealing with. This makes them vulnerable to influence (naively or corruptly) by those who do have the expertise or vested interest.

@ScottMGS @blacklight That, most of them being corrupt, and the fact it's usually about who has the most money to buy themselves a campaign.

@amberage @blacklight Or subservience to a party who'll provide the money to run.

@amberage elected officials restrained with the right legal framework for data governance are the lesser of the two evils.

After all, you don't need a lot of personal data to run a smart city. You may use video footage to train your models (for example, to recognize a garbage bag on a sidewalk, a bus, a bike or a pedestrian crossing at the wrong spot). But training those models is quite straightforward nowadays. Normally, you should throw away the footage data after you validate your models, and you also throw it away after feeding it to your live model and getting a prediction out of it, and you usually won't need it anymore.

Besides the use-case of training CV models, you may not need to hoard PIIs at all in order to run a smart city. An optimal model for garbage collection is one that takes into account neighbourhood density, collects live data about things like capacity of the bins or garbage on the street, and schedules collection accordingly. An automated model for public transportation scheduling apparently needs to collect a lot of live information about travelers, but actually all it needs is to know how many checkins and checkouts occur in a given time frame, and generate a unique random hash that connects a checkin to a checkout - no need to collect the ID of the public transport card nor other PIIs.

In other words, the problem of building smart cities is mostly a problem of collecting accurate aggregate data that openly available algorithms can operate on. Such aggregate data shouldn't contain any personal references whatsoever, therefore there's nothing wrong in sharing it with researchers and the OSINT community, and the algorithms should be open as well.

If we could build such a framework for data governance (only collect the data that you need, for the problem that you need, and aggregate it on the dimensions that you need in order to make it basically anonymous), and the data was open to everyone, and the algorithms that take the decisions on that data were available to everyone as well (there isn't much machine learning behind many of those models, some old-fashioned expert systems still do the job), then we have defined a clear framework that prevents abuse by design, and we put elected officials in charge of it that have the same access to the data as any OSINT enthusiast guy on Twitter.

My guess is that, of course, Google didn't want any of these things. They invested $1B of their money into the project, therefore they would have had no interest in collecting only anonymized aggregate data and make it open to the public. It's not a profitable investment if they didn't connect people's Google accounts to it somehow, and feed everything into their money-making ads machine. That's probably why they backtracked after many people became aware of the privacy issue.

@blacklight You're right about what would be necessary, but politicians have this annoying habit of loving surveillance and being corrupt. I admire your optimism though.

@amberage well, Canada has just managed to push back Google's attempt of building a surveillance capitalism experiment on a urban scale. The EU luckily is managed by people like Vestager and her team who are very sensitive on the topics of user privacy and data collection, even in the public sphere.

If we put the right people with the right intentions in charge of writing the rules, and both businesses and regulators start to follow the critical mass, then it'll be very hard to change those rules.

@blacklight The EU is managed by people who are currently trying to implement AI-based, cause-less mass surveillance of every text message. And by people who have effectively made memes illegal and tried to charge people licensing fees for linking to other websites.

@blacklight as a Torontonians, I followed this unfolding and attended some public meetings with the project's rotation roster of privacy commissioners, each forced to resign in turn when they realized their job was not data protection design, but breach blame absolution.

The idea that any elected official could be in control of that process, and then trusted with the results, is naive.

'Let them sell out the public trust and then get replaced' is not a good model.

Another problem with a Eurocentric perspective (Americas included here!), AMGAF may well try in weaker jurisdictions, e.g. Cambridge Analytical scandal started in Kenya.
@blacklight It's too bad there's no going back to "Don't be evil".

@blacklight
Is 'Smart Cities' something we actually need/want or a solution in search of a problem.

Autonomous garbage collection? What's wrong with a fixed schedule where the garbage is picked up? That means the garbage is only put outside on collection days.

Public transport? Fixed schedules are really useful. And you don't need sensors everywhere to notice if/when there is a capacity problem.

We should stop with the data collection fetish, which is what Smart Cities is all about (imo).

@FreePietje@x0f.org @blacklight@social.platypush.tech it's very dangerous when it comes to public transport, empty bus doesn't mean it's not needed

@potato_lisper @FreePietje this comes up to the brains of the people who read and interpret the data. The data may tell you that a bus often travels empty, then it's up to the policymaker what to do with it - is it the only service that serves a certain area? does it stop in places where it usually collects no passengers at all? how much margin do I have to modify the route or the schedule without impacting a few daily commuters?

The data only tells you about a problem or a pattern, then it's up to policymakers to interpret that data and take the right decisions on the basis of their knowledge.

@FreePietje there's actually a lot of untapped potential in smart cities - they just don't have to turn into a dystopian surveillance nightmare in the process.

Speaking of the Netherlands, Amsterdam has many "smart traffic lights" that combine cameras and weight sensors under the road to tell how much traffic is on each of the roads, and therefore how the flow should be managed to minimize the waiting time. Plus, historical data can be collected and analyzed offline to come up with better urban planning - for example, by analyzing the most common bottlenecks, roads that sustain peaks around specific times of the day, roads where pedestrians have to wait too long, etc. The thing is that you have to collect *zero* personal data in order to train these models.

Seoul is another good example of how to build a smart city that makes citizens' lives better, while respecting their data: cta.tech/Resources/i3-Magazine. The Mayor's office gets live data about the prices at every market in the city, so they can literally monitor inflation in realtime. They also designed the Owl Bus by analyzing data from several sources in order to estimate the areas of the city with highest demand in night time - all without revealing users' data: all we care is to get aggregate numbers, not to know who was there at a specific time.

Urban planning can definitely benefit from these technologies. The more accurate and timely data is fed to policymakers, the more likely they are to take decisions that solve the actual problems.

We just need to make sure that personal data is stripped unless strictly required (and motivated), and that the data and the code are both open so that OSINT enthusiasts and policymakers have the same tools to validate the decisions.

@blacklight
The "smart traffic lights" (TL) in the sense of the sensors under the road to detect whether there is traffic that needs to pass the TL, is technology done right AFAIC.
I was not aware of cameras and wonder whether that's actually needed. I'll leave that part out.

Previously the TLs turned green (and others red) at a fixed schedule, regardless whether there was traffic to be served or not. So people were waiting for traffic that wasn't there.

So here sensors are used to fix an ...

@blacklight
actual problem, which only detects that there is traffic, but doesn't know what/who/etc as that's not needed. It also doesn't need to store data.

Smart city initiatives as suggested by tech cos (not just Google) want to apply max tech with max data collection as possible (Eindhoven f.e.: x0f.org/@FreePietje/1016082443, they also wanted to add Facial Recognition to it). And then they'll look what potential problems could be solved *by tech*.

I think that 'tech' is far too often used, ...

@blacklight
where common sense and low/no tech solutions would suffice. The person emptying the garbage bins will notice when certain bins are often (over-)full and can report that so that extra/larger bins will be placed there. No need for high-tech sensors for that.

In x0f.org/@FreePietje/1016198504 you can read (in Dutch) about the CTO's of A'dam and Barcelona where they do it right, but those women don't look for a tech solution, but for a solution where optionally tech could be useful. ...

@blacklight
And then strive for minimal data collection. My fear is that in too many places/cities the required knowledge/skepticism is lacking and are (too) sensitive to tech solutions, which primarily benefit tech cos, not the citizens.

With the right people in the right places it could be useful (although it looks like I see less 'opportunities' then you do), but the risks of a dystopian surveillance nightmare is FAR too high imo.

@FreePietje we agree on the point that smart cities can be built with minimal data collection, and in almost none of the cases one needs to collect PIIs in order to run a smart city.

And this is exactly the point that policymakers need to stress as much as possible and enshrine into law if required. Data collection must be justified by a clear use-case for that data, while corporations like much more the "store now, figure out later what to do with it" approach - because, in the meantime that people figure out what to do with the harvested data, those corporation can make a big profit out of it.

Smart cities really have a lot of opportunities, and as more and more local governments turn to them we need to make sure that a clear legal framework is defined. It's not sufficient to say "we don't need this stuff", or "we can do it without tech". We should instead clearly define the rules of the game, and then pick the willing instead of the winners. Otherwise it's just a matter of time before Google picks another municipality that is a bit less worried about data management and they roll out their dystopian plan over there.

@blacklight
You have much more confidence that policy makers will do the right thing, then I do.

Take Munich f.e. itsfoss.com/munich-linux-failu where it took one (corrupt?) politician to kill an amazing initiative.

Or EU/MEP who should have (the resources to) access to all the right info, but I'm regularly dumbfounded by the things they propose.
Local policy makers don't have that access and are therefor (much) more vulnerable to the BullShit Bingo Card (Blockchain/ML/etc) and think that's 'smart'.

@blacklight
Plus what I said earlier that we (imo) too often look for tech solutions. So much so that (it appears that) we are controlled by tech instead of we using tech as a tool for specific purposes.

Humans have an enormous capacity to come up with creative solutions and it seems we are forgetting that.

@FreePietje the problem is that policymakers can't properly solve a problem if they don't have enough data to solve it - and that's where smart cities come in, and that's also like e.g. the local government is using these technologies in Seoul.

Of course, politicians will be politicians, and a dumb one or a corrupt one put in charge of things is enough to ruin everything. But it's sad to give up on the idea of using better technologies to improve our lives just because those in charge are too dumb or corrupt to implement them properly.

@blacklight
Policy makers and urban planning are involved in where and how to go to for the next 10/20/50 years. Like f.e. making a city center car free.
That involves vision and knowledge on how to achieve that, not live data.

I just read the article you linked about Seoul and all I see are items from the BSBC (Big Data/IoT/5G) and solutions in search of a problem. CTA isn't exactly an impartial source; they have tech 'solutions' and are looking for problems/consumers to buy them.

That ...

@blacklight
S-Korea wants to promote 'Smart Cities' (by making Seoul one) is understandable from an economic standpoint: they make a lot of the tech.

Frightening example: they tracked all people's location data to determine where the night buses needed to go? IMO you need to make them go to places where (young) ppl live (city knows that) and then at strategic points in those areas.
And make it predictable so ppl can rely on them. Dynamically adjusting them on live data makes them unreliable.

@blacklight ...and they never would have addressed it, as the whole premise was to catch as much telemetry and proximal into on everyone there, in perptuity.

Sign in to participate in the conversation
Mastodon

A platform about automation, open-source, software development, data science, science and tech.