There is interesting news out of MIT. A team used synonyms for words to attack NLP (natural language processing) systems and reduce the quality of results. They tested this on Google’s BERT system to show how their app (TextFooler) can expose areas that need improvement. NLP systems are at the core of Google, Siri, and, Alexa so efforts like this can help them to get better.
The academic paper can be read/downloaded from THIS LINK
“…tiny alterations to the input can flummox an AI and make it misclassify what it sees. TextFooler shows that this style of attack also breaks NLP, the AI behind virtual assistants—such as Siri, Alexa and Google Home—as well as other language classifiers like spam filters and hate-speech detectors” Full post is available at TechnologyReview.com