GPT Models: Trustworthy Companions or Mischievous Tricksters?