Volunteers couldn’t tell AI-generated comments from those penned by humans. Read more at Ars Technica…