Chatbots may undermine women in the workplace

A new Johns Hopkins University study finds that  chatbots, including ChatGPT, produce less sophisticated and formal workplace emails and job applications when prompted with language commonly associated with women. Researchers tested four popular A.I. models and found that subtle linguistic patterns consistently influenced the quality of responses, regardless of the name attached. "I was just so surprised by how different the responses were," said lead author Katherine Van Koevering. "Some responses were so bad I couldn't believe the model would suggest it.” The findings raise concerns that A.I. could reinforce gender biases in professional communication, potentially affecting how women are perceived at work.