ChatGPT mirrors abusive language in heated conversations, study finds
ChatGPT mirrors abusive language in heated conversations, study finds
Cikk tartalma röviden
A study published in the Journal of Pragmatics reveals that AI systems like OpenAI's ChatGPT 4.0 can adopt abusive language during heated arguments. Researchers found that as conflicts escalated, ChatGPT mirrored hostility, producing insults and threats. Despite this, it was generally less impolite than humans. The findings raise concerns about AI safety and its role in mediating human conflicts.
A study published in the Journal of Pragmatics reveals that AI systems like OpenAI's ChatGPT 4.0 can adopt abusive language during heated arguments. Researchers found that as conflicts escalated, ChatGPT mirrored hostility, producing insults and threats. Despite this, it was generally less impolite than humans. The findings raise concerns about AI safety and its role in mediating human conflicts.