A Twitter botnet has been discovered that uses ChatGPT to promote fraudulent cryptocurrency schemes and websites. 

Comprising three news sites, one named 'Fox8', and a cluster of 1,140 Twitter accounts, the botnet was discovered by Indiana University researchers searching Twitter for the phrase 'as an ai language model’ - a common phrase generated by ChatGPT when it receives a prompt that violates its usage policies - between October 2022 and April 2023.

The bots create and manage fake accounts and generate human-like tweets to promote dubious crypto investments and have been implicated in attempts to steal from existing crypto wallets. The bots strategically engage with legitimate accounts, enhancing their credibility and making it challenging for users to differentiate between authentic and malicious content.

The botnet poses significant financial risks to unsuspecting investors, potentially leading to substantial monetary losses. 

The discovery also raises questions about OpenAI's ability or willingness to detect and address the misuse of ChatGPT and its other AI systems.