On Monday, March 27, Europol's Innovation Lab warned that cybercriminals are working to exploit the ever-evolving capabilities of the GPT Chat chatbot to enhance their fraudulent operations, such as phishing campaigns. The European agency also mentions disinformation capacities, facilitating and enhancing the narrative to impersonate authority figures and institutions close to reality.
More worryingly, the LLM (Large Language Models) chatbot is very useful for non-experts to democratize malware coding, presaging an increase in attacks.
GPT chat safeguards, the security measures in place, are considered easy to compromise, but they are still in their infancy and shall improve. Last week there was an incident where sensitive user data was leaked, with the same problem as sharing chat history headlines.
Download the full Europol’s document here: https://www.europol.europa.eu/
