| From the article (mistakes highlighted): > Human news writers regularly point out that AIs tend to lack nuance and a _flare_ for language in the stories they churn out. That’s probably a _fare_ criticism [...] Maybe they used speech-to-text transcription for this, given that the mistakes are homophones? It seems very unlikely that either a human typing this, or a computerized system would make these mistakes (if it learns word associations from a corpus). PS: the article also claims to be human generated: > This story was not generated by an AI, but to be fair, I haven’t had my coffee yet. EDIT: Oops, I might have misunderstood which article you were referring to, since the reference was not placed next to "this". |
You underestimate people's ability to make language errors, including spelling ones. Every time I see somebody I suspect is a native English speaher using "it's" for "its", I grind my teeth. (Another instance is somebody using phrase like "as a programmer, the data bus should be written..." to mean "I, as a programmer, think that..."; this phrasing makes me simply furious.) With those errors they make reading my second language so much harder, and I can't even point their bad spelling or writing style out, because I'm seen as being nitpicking or something.