The article discusses the persistent threat of social engineering email attacks, necessitating a shift towards automated identification of weak explainable phishing indicators (WEPI) using NLP to supplement user vigilance, as highlighted by an annotated corpus of 940 emails with 32 WEPI labels; this approach focuses on adapting anti-phishing training for both humans and machines by detailing indicators like urgency and verifiable mismatches, aiming to develop models that ease cognitive burdens and promote combined human-machine approaches for enhanced phishing detection. ```

 Using NLP to identify weak explainable phishing indicators (WEPI) in emails