As the article points out, the capability to scan things on devices is already deployed on huge swathes of devices in the from of virus scanners and content search indexing. All the photos on iPhones and many Android devices get scanned for text recognition now anyway. I'm sorry, but 'OMG scanning things" is a genie that got out of the bottle decades ago. So no, scanning content is not a dangerous new capability, it's an extensions of existing capabilities to a new use case.
Now arguably this is a further slip down that slope of on-device scanning, but it is not some new watershed moment where scanning is happening for the first time. Also we can see that the trend is not to repurpose existing scanning systems for new purposes. Virus scanning, search indexing, text recognition and now this are all separate mechanisms implemented and operating independently in task specific ways. The prediction that these mechanisms will be subverted for other purposes has not panned out.
The real issue is what is being scanned and why, and the implementation of this system. That's the topic we need to focus on.
> As the article points out, the capability to scan things on devices is already deployed on huge swathes of devices in the from of virus scanners and content search indexing.
that's a ridiculous counterpoint. 1) no iphone has AV 2) AV and content indexing don't secretly contact the feds.
adding a new "background service that sometimes contacts the US government" is a big privacy regression.
As the article points out, the capability to scan things on devices is already deployed on huge swathes of devices in the from of virus scanners and content search indexing. All the photos on iPhones and many Android devices get scanned for text recognition now anyway. I'm sorry, but 'OMG scanning things" is a genie that got out of the bottle decades ago. So no, scanning content is not a dangerous new capability, it's an extensions of existing capabilities to a new use case.
Now arguably this is a further slip down that slope of on-device scanning, but it is not some new watershed moment where scanning is happening for the first time. Also we can see that the trend is not to repurpose existing scanning systems for new purposes. Virus scanning, search indexing, text recognition and now this are all separate mechanisms implemented and operating independently in task specific ways. The prediction that these mechanisms will be subverted for other purposes has not panned out.
The real issue is what is being scanned and why, and the implementation of this system. That's the topic we need to focus on.