Clearview AI: comment from us in I News
Clearview is a controversial US facial recognition company raising concerns over privacy after it created a database using nearly three billion public photos of people from sites such as Facebook, Twitter and YouTube.
Andy was asked for comment and insight by I News: published article here.
His lengthier and unpublished insights below:
1.) What does Clearview highlight about facial recognition and social media?
Taking at "face" value Clearview's claim that they do search any private or protected info in social media accounts, this raises thorny questions about privacy, security and capitalism. The issue is that just because something is posted in a “public” social media space, does not make it fair game to re-use without informed consent.
It should be noted too that Clearview are not alone in using public social media data without consent. For example, many companies use “public” social media data to train their algorithms (that in turn are used to sense emotion expressions in faces).
2.) What are the concerns for the UK when looking at its mining of faces using social media, very often without permission, but is still being used by law enforcement?
The question at heart is an old one: it is the Hobbes debate revisited, but with a twist. We typically talk about the balance between privacy and security, but for Hobbes liberty means giving up some rights to establish peace via an authority that works on our behalf for the common interest (such as with policing). In exchange, people obtain protection, providing they obey the law. In other words, we have a social contract. There are numerous qualifiers to this, not least that citizens are also meant to be happy as well as protected. For Hobbes, Safety cannot come at any cost or be the sole concern of life in society.
What Hobbes did not have in mind (and here’s the twist) is that it would be private business providing the security. This is a very different social arrangement and contract. Given the above concerns about data shared for one reason being appropriated for another reason, and the nature of the Hobbes trade-off between Government and citizens, as an absolute minimum this needs civic debate and parliamentary scrutiny. In no way should this be installed in the UK without scrutiny. I would hope the ICO would take a strong view on it too.
3.) Could this feasibly be introduced here to create a database for law enforcement?
This has to be seen in context of the willingness to use facial recognition technologies without adequate societal conversation about whether this is at all desirable. I think it is an unnecessary intrusion into peoples’ daily lives, but more objectively it is based on commercial use of personal and arguably sensitive data that citizens have not consented to. If police were keen, I would hope the ICO would take a strong view and the courts would block such usage.