We don’t need AI videos of fake animals. There are real ones out there and they’re really cute | Rebecca Shaw
TL;DR
Rebecca Shaw uses her Guardian column for a blunt anti-GenAI argument: the everyday use of generated images, text and videos has already gone too far. Her core example is cute animal videos: once AI fakes flood the feed, even real nature clips start to look suspicious and lose their tiny burst of joy. Shaw ties the cultural frustration to environmental criticism, especially water-hungry datacentres and throwaway prompts for trivial tasks.
Nauti's Take
Shaw’s argument is sharp, but the core holds: AI content does not just consume compute, it eats context. A real clip now has to fight the suspicion that it was generated.
That is the ugly side effect of a tool culture that wants to synthesize every little moment of wonder. The practical lesson is simple: anyone publishing with AI needs labeling, provenance and restraint.
Otherwise more content just means more distrust.
Briefingshow
The column lands on a problem many AI debates frame too technically: trust is a product feature. If users inspect every clip for artifacts before enjoying it, platforms lose authenticity and casual delight at the same time. For AI companies, that is an adoption problem a better video model cannot fix by itself.
Sources