What's your point? The areal cameras in the shop, regardless of zoom level, are not being replaced by the cameras in your face and only in your face; it's an additive situation.

The cameras in your face are only effective for capturing your face/person in detail, and they are presumably being applied to every customer wholesale, automatically during checkout, and probably associated with your purchase and payment information.

You're arguing as if this isn't any different than the status quo, when it obviously is.

Now we not only have the areal cameras that were presumably only zoomed-in on suspicious behaving customers for loss prevention, we also have every single in-person purchase being captured zoomed-in, regardless of behavior, with a camera lens right in your face.

It's as if retail has decided to actively encourage everyone with any shred of self-respect to stop shopping retail and buy everything from Amazon. If you treat me like a presumed criminal I'm going to take my business elsewhere. What's next, metal detectors and armed guards at the doors?

Installation of cameras at face level is nothing new. Equipment like self checkout kiosks and ATMs have had (usually discrete) cameras in the user's eyeline for decades. Higher theft targets like banks and convenience stores install (often hidden, sometimes not) cameras at eye level at the entrance and exit doors. Hidden in the height stick is common in banks, grocery stores tend to just install normal domes at head height. This is all in response to the simple fact that wearing a baseball cap is enough to prevent ceiling-mounted cameras getting your face, and your face is exactly what the police want to be able to see after a theft.

I don't intend to defend the use of facial recognition in these applications, just saying that cameras for face capture date back to the early days of continuous video surveillance in the '80s, and although I haven't seen it myself I assume there were some users that installed hold-up cameras at eye level in the '60s - certainly the importance of getting a clear view of the face was discussed in articles about hold up cameras, but in that time it was less common for people to wear hats indoors (and bank security would typically require people to remove them).

Here's the thing about the entire argument thread.

It not being new is not reason enough for it to continue.

It already existing does not mean we should never revisit this decision.

The technology on the other side of the feed is new and it has changed tremendously.

Has it, though? While much discussed, facial recognition on surveillance is very expensive in practice and so uncommon to a degree I think many here would find surprising. The article here discussed facial imagery captured by handheld cameras, not involving the surveillance system. That's why I think it's important to split these hairs: video surveillance recording of faces for post-incident use is very different from proactive use of facial recognition. The former has been accepted for decades as a practical necessity for investigation of crime. The latter is new. I think that framing the discussion around cameras at face level, which is not new, misses the point and starts from a premise that will be wholly rejected by the industry. We need to focus instead on the use of automated technology to process this data, which is new and not yet widely practiced.

That is changing. Newer self checkout terminals correlate video from both cameras with purchase history, primarily as a short-term local aid to staff when figuring out discrepancies. But this can obviously be extended to a long-term retention business intelligence application. The time is now to put pressure against this kind of data collection. But saying "there should be no cameras pointed at the face" will not apply that pressure. It will be dismissed because the use of cameras for face capture is a universally accepted anti-crime practice, and arguing against it as a way to object to mass surveillance ignores the differences between short term retention of data for post incident use and mass collection.