By Professor Fraser Sampson, former UK Biometrics & Surveillance Camera Commissioner
Fears about AI-enabled biometric tools like facial recognition are often fuelled by what one legal researcher called omniveillance - the ability to watch all of us, all of the time. There's another term for something that covers the entire population and it's very topical again in the UK.
Since 2019, the word 'pandemic' has marked the moment when the latest variant of the corona virus was detected but the adjective comes from the Greek pan demos meaning 'across all people'. The sheer breadth and scale of the threat affected all the people and called for a pan demic response. Despite all the fictional portents, the policing scenario demanding the greatest level of surveillance saturation for the past 50 years came, not from crime and disorder, but from disease. Severe restrictions on movement and basic social interaction led to extensive surveillance and use of policing powers like stop and search and the issuing of criminal penalties.
In its first 'module' report, the UK COVID-19 inquiry has concluded that preparedness was inadequate, risk assessment flawed and lessons from previous civil emergencies had not been fully learned. There are several policing and law enforcement lessons we can learn from this one.
The global exigencies of the COVID-19 pandemic required temporary, emergency measures some of which involved segregation, restriction on movement, assembly and attendance at school, as well as the ban on public worship. Stopping people doing things that for the past half a century had been cherished constitutional freedoms was never going to be straightforward and an effective police response required a high level of vigilance over the entire population. This was a period of literal pan demic surveillance in both senses of the term, watching everyone to constrain a global disease. But even viewed against the scale of the emergency, the extraordinary powers for policing were seen by some as unacceptably broad in their scope and discriminatory in their execution. Using surveillance capability like drones and automated number plate recognition (ANPR), local police forces became agencies of health enforcement. A surge in citizen stops, chaotic communication and - in a move away from the Peelian principles - the overzealous issuing of on-the-spot fines, led to allegations that the response undermined the British model of policing by consent.
Some lawyers will say that this is just another question for proportionality - pandemic threats warrant pan demic solutions - but does every global emergency threatening the lives of millions justify near total state surveillance? Many scientists have testified to the immediate and irreversible threat to humanity from global warming and climate change. If that's not pan demos I don't know what is, but does it mean local government emission zone surveillance can be as extensive and intrusive as that for COVID-19? When its purpose is to avoid the death of our planet, what would a dis-proportionate use of AI-enabled surveillance look like?
The inquiry noted how tackling the spread of the virus across the population relied on the enlistment of the population and suggests more volunteers will need to be 'engaged' in future crisis responses. This will bring some interesting surveillance issues. As I've noted, public space surveillance is no longer about where the police put their cameras, it's about what they do with the data from everyone's cameras. Since 2019, there's been a big uplift in both domestic surveillance capability and our willingness to put it at the disposal of the police in an emergency. We are tracking everything from pets to golf balls, cheaply and reliably and if citizens make the content of their private devices available, why wouldn't the police use it to save lives? How could they decide which data to ignore? In tackling future health crises, our aggregated surveillance capability will be a powerful community asset. But self-surveillance is far more intrusive than anything the police operate directly and the lines around processing our personal data are much more restrictive for law enforcement than those constraining what we do domestically. When lending our fine location apps watching our children and smart device captures from doorbells and dashcams to reinforce the mass surveillance capability of the state, we enter a very grey area, legally and societally.
The inquiry recommends a new UK-wide approach to the development of better systems of data collection and sharing. Advances in biometrics and surveillance have put our capability and capacity to prepare for, respond to and recover from global crises beyond even what was available only six years ago; at the same time, this has created dependencies and vulnerabilities on a similar scale. At times of universal threat, pragmatism rules - switching off the constitutional firewall is easy, but can all the AI watchdogs be put back in the kennel once the emergency has passed? Levitsky and Ziblatt observe that during security crises, "citizens are more likely to tolerate and even support authoritarian measures - especially when they fear for their own safety." Next time, AI-enabled capabilities combing databases and combining results with Global Positioning Satellite (GPS) information, vehicle telemetrics and Frankenstein CCTV cameras will give governments a formidable weapon to discharge their overarching duty - to keep their people safe. But it will be crucial to ensure that any pan demic response is transient and used only to the extent necessary to counter the pandemic threat. Who will we trust to monitor and enforce that?
We have probably not witnessed the last global-level infection and the next one will be monitored by AI-enabled technology. The coalescence of two critical features - the need to monitor the entire population and the technological ability to achieve it - makes it a practical certainty. In preparing for it, we should start modelling the realities of pan demic surveillance now.
Fraser Sampson, former UK Biometrics & Surveillance Camera Commissioner, is Professor of Governance and National Security at CENTRIC (Centre for Excellence in Terrorism, Resilience, Intelligence & Organised Crime Research) and a non-executive director at Facewatch.