Who watches the watchmen? Why Europe is right to ban facial recognition
Tue 4 Feb 2020 | Robert Brown
Applying the brakes on facial recognition technology is a smart move, argues Cognizant’s Robert Brown
The potential five-year ban on facial recognition technology in public areas by the European Commission and the condemnation by civil liberties groups of the introduction of facial recognition by the Met Police in the UK is not surprising. The health of our democracy demands trust, and the tech trust deficit is not closing. Too often, there is retroactive regret that not enough was done to prevent something from happening.
Meanwhile, there is near-unanimous agreement that flaws with facial recognition software are unacceptable. The error rates, especially for people of colour, are far too high for it to be used in any law enforcement frame for the general public. We need to close the gap in terms of the percentage of error rates.
Bans on facial recognition, such as those suggested by the European Commission, raise valid concerns over the need to walk before we run. From a UK perspective, there is a heritage of common law and civil law, as is the case across most of Europe. Given most countries are bound by principles of being innocent until proven guilty, those error rates among minority communities are unacceptably high. There needs to be a veto on these kinds of technologies until we can have a very high degree of certainty.
Companies developing facial recognition technology need to undertake much more performance testing before it is ready to go live.
As part of developing further safeguards around facial recognition technology, one role Cognizant expects to emerge is the ‘algorithm bias auditor’, the aim of which will be to make sure that any bias is eliminated from a company’s technology before it goes live.
If the ban were to be voted in by the EU Member States, the European Commission will require that any facial recognition technology implemented in public places not be overshadowed by ‘the black box problem’, where we need to be able to fully understand the technology’s algorithms and how it is working. As such, the algorithm bias auditor role will be fundamental to every business’s competitiveness. Until these kinds of roles are put in place, facial recognition and other emerging technologies will remain under heavy scrutiny from the regulators.
When it comes to regulating data – which underpins facial recognition – Europe is in the lead: GDPR alone was the first real data regulation anywhere in the world.
Looking further afield, the use of facial recognition in other countries can highlight why regulation is no bad thing. For example, China has more facial recognition cameras than anywhere else in the world, and the population is showing growing concern over how much their movements may be being monitored. Facial recognition can now be used for payments, showing the technology’s potential; however, China’s full-blown commitment to its deployment has helped highlight many issues surrounding state surveillance.
In the US, different communities are undecided over whether to adopt facial recognition or regulate against it. Meanwhile, one of the great ironies is that San Francisco, arguably the epicentre of all things tech, banned facial recognition last year.
While a lot, if not most, of the research and development happening in facial recognition technology is taking place in Silicon Valley, there is no coherent or consistent US policy view of how to get a grip on this thorny issue. This ultimately means that Europe is still in pole position when it comes to adopting facial recognition technology, and that applying the brakes could ultimately prove its smartest move.