I fear it would be a pyrrhic victory at best; all it takes is one instance of acceptance (via smartphone update, or an infinite number of other avenues) for it to propagate to every other entity.
That’s actually before encountering ownership issues of photos, as it usually is the photographer who owns the copyright to an image - and if they upload that photo to a service and agree for it to be trained upon; what happens next?
I fear it would be a pyrrhic victory at best; all it takes is one instance of acceptance (via smartphone update, or an infinite number of other avenues) for it to propagate to every other entity.
That’s actually before encountering ownership issues of photos, as it usually is the photographer who owns the copyright to an image - and if they upload that photo to a service and agree for it to be trained upon; what happens next?
I think you might be overestimating cooperation between these companies, but it’s definitely a valid concern.
Probably, yes - but in a race for as much data as possible to try and feed their LLM models, I wouldn’t put it past them.