Has recent controversy scuppered facial recognition’s chances of being a key solution for smart cities? Kate O’Flaherty reports.
It has the ability to enable multiple applications that keep smart cities safe, but as facial recognition technology has become more sophisticated, so has its potential for abuse. It suffers from problems with accuracy and racial and gender bias, and privacy concerns have led to widespread pushback against the technology.
The cliché is that it could be a modern-day version of Jeremy Bentham’s Panopticon – the prison where the warden could see every prisoner and cell, but the inmates couldn’t see their jailer. However, it is in the justice context where facial recognition is causing the most concern. Over the last few months, the killing of George Floyd in the United States has brought police powers, and their use of facial recognition, to the fore around the world.
Facial recognition trials in the UK have been met with opposition. Elsewhere in the United States, San Francisco was the first city to implement an outright ban, and lobbying groups in Chicago have called for a halt to police use of facial recognition technologies. In response to the backlash, major players including Amazon, Microsoft and IBM have pulled back on development of facial recognition technology.
Even before the big tech firms pulled back, facial recognition was already facing regulatory challenges. The US lacks adequate regulation to handle the problems with the technology and earlier this year, the EU mooted a five year ban of facial recognition after finding it is prone to inaccuracy, can be used to breach privacy laws, and can facilitate identity fraud.
Between criticism and barriers to entry, will facial recognition fail to take its place among the technologies of a modern smart city? The technology’s roll out now depends on developing regulation and standards that will help protect the privacy of citizens. Zak Doffman, CEO at surveillance solutions company Digital Barriers says it’s “critical the industry, as well as lawmakers and society at large” agree how to strike a balance between public safety and ensuring the technology is not misused.
The impact of the facial recognition backlash on smart cities is not straightforward. Other perhaps lesser known technology providers offer applications that might be considered more ethical.
As Doffman points out, not all facial recognition systems are the same. “Even as we push back on so-called standoff surveillance applications – those that have made recent headlines – we are seeing more use of facial recognition to make travel more efficient and secure. The biometric ePassport gates at airports, for example.”
Covid-19 has also raised the need for additional facial recognition use cases. Doffman cites the example of contact-free identity assurance — a key part of smart city deployments that he says, so far, hasn’t seen much push back. “The idea that a ticket or a pass might shift from a physical ID to a smart device recognising my face clearly negates me signing in or carrying and handing over a physical card or document. We are also seeing facial recognition becoming the norm to unlock smartphones and for some other forms of access control.”
At the same time, despite privacy concerns, the technology isn’t hated across the board. In fact it’s often accepted for public safety applications, as long as it’s accurate. “When it comes to public security, citizens recognise the benefits of facial recognition technology,” says Pierre-Adrien Hanania, global offer leader, AI in public sector, Capgemini.
Two thirds of people are comfortable with the use of AI-enabled cameras capable of detecting and tracking abnormal or alarming situations in public areas, and over half see the benefit of facial recognition technologies to track offenders, according to a recent Capgemini Research Institute report.
Many other dimensions of pattern and shape recognition are already being leveraged in smart city initiatives, says Hanania. “For example, the technology can help recognise garbage on the side of the streets, or broken infrastructure which could harm citizens and needs to be maintained or replaced.” And despite the pullback by tech firms, a July paper by Global Market Insights predicts the market valuation of facial recognition will surpass $12 billion by 2026 from $3.4 billion in 2019.
Much depends on companies being able to implement the technology in a way that considers the implications for those residing in the smart city environment. AI software company NtechLab CEO Artem Kukharenko accepts there are ethical considerations, but says his firm has trained its software’s neural networks with stimuli from all races, skin tones, and genders in equal proportions. “Concerns have been raised around misidentification and bias, and we are aware of the issues that several facial recognition systems have displayed.”
But some industry experts don’t even think facial recognition is that key to smart cities’ future. In some cases, they say, this technology can be replaced — at least for now. For smart cities to improve, says Alan Marcus, chief digital strategy officer, Planet Smart City, it is helpful to know where people congregate, as well as male to female ratios, for example. “So how did we do that in the past? People would just stand on the street and watch. And do I need to know the actual identity of people in most cases? There are other options that can be used for basic smart services…Having facial recognition isn’t an essential factor.”
Future facial recognition in smart cities
Despite recent controversy, it doesn’t seem that facial recognition will be brought to a sudden and permanent halt. Much of it is about ensuring companies use the technology properly and for its intended purposes only. “That is a hard choice — unless policies, governance, and regulation are in place to manage it,” Andy Thurai, industry analyst at GigaOm says.
Taking into account that these factors are lacking especially in the US, law enforcement applications such as transit monitoring or cameras to identify suspects have the potential to be “inhibited in the short term,” says Forrester VP and research director Merritt Maxim.
Also in the short term, smart cities in some countries will be more affected than others. The US and Europe are pushing back on the use of facial recognition within law enforcement – and that’s “very different in Asia and parts of the Middle East, for example”, says Doffman.
Philip Ingram, MBE, a former colonel in British military intelligence thinks the decision to pull back may impact the roll out of smart city technologies in the US for some time. However, globally he thinks it will have little impact. “There are many other facial recognition application and smart city capability providers and IBM and Amazon have only restricted their sales to the US. Of note…Huawei is a global provider and other Chinese companies such as HikVision (the world’s largest CCTV manufacturer) have the ability to meet wider global needs with advanced technologies.”
The issue of facial recognition in smart cities is challenging to predict as it covers many possible applications, says applied futurist Tom Cheesewright. “Networked CCTV with facial recognition on a city-wide scale presents a fairly intrusive threat to people’s privacy. But closed loop facial recognition for entry control, for example, could offer people enhanced security and a much more low friction experience.”
For now, smart cities will just have to wait and see what happens, and in the meantime make use of the technology they already have. It’s very hard to predict, but Maxim thinks the situation in the US may change if public opinion shifts. “The larger vendors will do the right thing, but they will vote with their wallets and if they see the business opportunity is something they cannot ignore, it will be hard to resist longer term,” he says.
But he warns: “If they do change their minds, they will need to be clear why. They also need to show that facial recognition won’t abuse consumer rights and be prone to racial bias and gender bias – the ethics of AI will be a key part of reassuring customers.”