It was inevitable that recruiters would be amongst the first to use GAI for candidate outreach and equally inevitable that the most likely target of that outreach - software engineers - would be first to develop counter measures against it. New phrase for us to learn: “
prompt injection attack”- a technique where you place an instruction into the data which you think the AI will ingest, in order to corrupt the output so that you, the human candidate, can readily identify a generated message. I’m sure the fellow here was just experimenting but a great illustration as to how this class of technology works, and of human ingenuity in making sure it doesn’t