I think the tweet misses the point of these demonstrations. AI has the potential of being a dangerous technology and it's good that people are showing how it can be fooled into doing things it shouldn't.