CEO Suggests That Humans Could Be "Adversarially Attacked" Like Neural Networks
Briefly

There could be, an artificially constructed sensory input that makes you go insane forever. - Tomáš Daniš
It seems 'completely obvious' that such an attack - known as a 'jailbreak' in AI - could be used on humans as well. - Florent Crivello
Crivello referenced a 2015 Google study where overlaying a noise-laden image on a panda photo made a neural network misidentify it. This concept can be applied to humans too. - Article
Crivello highlighted the 'Pokémon' episode inducing seizures in Japan, demonstrating the harmful impact of visual stimuli. This could set a precedent for using simple imagery to harm humans. - Article
Read at Futurism
[
|
]