Why do you think the cameras can be turned back on?
‘After the uproar that arose in 2021 following Mare’s reporting, we started two trajectories: a security trajectory and a privacy trajectory. In terms of security, five risks emerged which we have since addressed. We checked whether the measures were sufficient, and it turned out that they were.

'In March 2022, we conducted another public test, in which we let hackers and students attempt a final security breach. The results showed that security is now safeguarded and the problems brought to light by Mare have been resolved.

‘The privacy investigation took a little longer because first, the data protection officer had to draw up a report from his independent position. His main recommendation was to carry out a data protection impact assessment (DPIA, Ed.). Or rather: that we should have done so already. We had an independent party carry out the DPIA and we asked experts in our own community to have a look at it and add to the report.’

What does the DPIA show?
‘That several of the risks have been resolved. Some of the risks still need to be addressed, but we will do so before turning the scanners on. For example, making agreements with the supplier about the sensor master key and no further processing of personal data.'

‘There is also a residual risk which we’re not going to fix, let me be clear about that. When the buildings are closed in the evenings and at night, you’re supposed to turn the scanners off. That has a number of implications, however. To turn the devices off, you have to cut the power to the network, which means other devices on the network also turn off. In that case there is the risk the sensors lose their calibration or don’t reboot automatically. That would mean recalibrating them the next day, which involves all sorts of potential risks. This was also indicated in the DPIA as a residual risk that is not insurmountable.

‘Therefore, we will leave them on during the night. That means the sensors will count people who are present in the evening or at night, such as cleaners or security guards. That residual risk is indicated in the DPIA and is smaller than the risk that comes with turning the cameras on and off every day.'

What happens to those night data?
‘We don’t use those. We only use the data to look at flows of people during the day. It is, of course, not our intention to start tracking persons individually for the purpose of taking measures.’

But it will, in fact, still be possible to track those few individuals at night, placing people with precarious jobs under even more scrutiny. You can’t just use camera surveillance either, so why is this allowed?
‘That’s the trade-off we have to make. I get what you’re saying, but on the other hand, the count is always anonymous. And even if you wanted to, you wouldn’t be able to confirm that a specific person walked somewhere based on the count. ‘That’s not what this tool is for and it won’t be used for these purposes.’

We got this system for the purpose of counting, but it can do much more. Isn’t it much too advanced?
'We want to use it to measure crowding in buildings. This system is very good at that. All of its other functionalities have been deactivated.

'In the Covid-19 period, we wanted to know whether or not we could safely use our rooms, now tu. We’ve been manually counting occupancy for years. That's a very labour-intensive and expensive process and only gives you data on two specific points in the year. That means we can’t make course corrections.

'In particular, students’ need for study places is very high. This system allows us to take interim measures to scale down classrooms, for example, and turn them into study places. We can count people flows in the buildings in a technologically advanced, continuous way and respond accordingly.'

But then, wouldn't a more basic system have sufficed too? Perhaps it would be slightly less accurate, but it would have avoided privacy risks.
‘That’s open to doubt. I asked the University Services Department which other systems they had considered. As it turns out, other tools, such as Wi-Fi tracking, were also fraught with privacy and security issues.

‘I think that the set of measures we’re taking ensures that privacy is well and truly safeguarded. Just as well as for alternative systems, maybe even better, and it allows us to measure more.'

So why is it that the university did recognise the privacy problems of those other systems beforehand, but took a rather laid-back, not to say lax, approach to the problems of this system?
‘I wouldn’t put it like that. If I could turn back time, I would have asked for privacy and security assurances in a different way. Privacy is now safeguarded more effectively; not just for this system, but also for other pilots we’re running, and we’re taking more measures. This won’t be the last system where privacy and security are at odds with each other. That is a consideration we have to make as administrators: whether the benefits of a technology weigh up against privacy and security.’

What if the University Council’s advice is negative?
‘That’s impossible to say because we have to consider the grounds on which the advice is based. Because it is a sensitive issue, we want to make sure that the community can get behind the use of this tool through the University Council. I do think that the measures we have taken provide us with the optimal conditions to safeguard privacy and security.

'And it's not us, the administrators, who benefit from the capabilities of this system, but primarily the students and staff, as it allows them to use our rooms effectively. But if they issue a strong advice based on new grounds, we will take it very seriously. Ideally, I would like for us to agree on the matter.'

And if the advice is positive, when will the cameras be turned on?
‘If the University Council gives us the green light on 13 February, we will need at least another month to take a number of measures. So I expect they won’t be turned on before April.’

Also read: Smart cameras have no place on campus: take them down