Attackers are exploiting ChatGPT’s trust to sneak in gambling promotions and hijacked recommendations. https://t.co/guvF5mD4jx
ChatGPT Agent - earlier version of the model that won the IMO Gold medal https://t.co/Ae8ardy9Mp https://t.co/PRY1nc5RYl
ChatGPT Agent → ✅ Unlocked AMA, along with your bank details 😏 https://t.co/vwZxCr7StQ
OpenAI has introduced a new feature called ChatGPT Agent, which has been publicly showcased since July 17, 2025. Users have demonstrated the agent's capabilities, including creating animations of scientific processes such as fusion and DNA replication. The ChatGPT Agent is described as an advanced iteration related to the model that won the International Mathematical Olympiad (IMO) Gold medal. However, security concerns have emerged, with reports indicating that attackers are exploiting the trust users place in ChatGPT to promote gambling and hijack recommendations. Despite these risks, the ChatGPT Agent continues to attract user engagement and exploration.