Europol is concerned about ChatGPT’s ability to help hackers and fraudsters

The EU police have issued an official warning about the potential misuse of ChatGPT, an artificial intelligence-based chatbot. Law enforcers note that AI can be used in phishing attempts, information and psychological operations, spreading disinformation, and facilitating cybercrime.

Since ChatGPT became available to all users, OpenAI’s competitors have been spending a lot of resources to launch similar products or integrate similar technologies into their own developments. Europol does not rule out that cybercriminals are already using this to create malware.

«As the capabilities of LLMs (large language models) such as ChatGPT are being actively improved, the potential use of these types of artificial intelligence systems by criminals poses a grim prospect,» Europol said during the presentation of its first AI technology report.

The document highlights the malicious use of ChatGPT in four criminal areas:

• ChatGPT’s ability to produce text that is not easily distinguishable from that of a human author makes it a useful tool for phishing.
• Due to its ability to reproduce speech patterns to mimic the speech style of specific individuals or groups, the chatbot can be used by criminals to gain trust in order to cause harm.

ChatGPT’s ability to quickly produce authentic text also makes it an ideal tool for propaganda and disinformation. «It allows users to generate and distribute messages that reflect a particular narrative with relatively little effort,» Europol emphasized.

• Criminals with limited technical knowledge can also use ChatGPT to create malicious code.

Since the world’s most popular chatbot has already been subject to global failures, Europol’s concerns make sense. As a reminder, ChatGPT recently published 1.2% of its users' payment data by mistake. OpenAI said that the leak was caused by a bug in the open source Redis service.