The cybersecurity experts on the Tenable Research team continuously monitor the threat landscape and examine the shifts in technologies and techniques employed by attackers. The work involves analyzing enormous amounts of data in order to provide contextually relevant guidance to the organizations using our products.
In recent months, the security community — as well as government agencies worldwide
— have raised concerns about the use of AI technologies like OpenAI’s ChatGPT. While the organizations developing these tools have demonstrated impressive advancements in generative AI, many in the security community recommend proceeding with caution.
These tools, built on LLMs, are evolving at breakneck speed. With their ability to produce human-like responses, LLMs provide ample opportunities for unique use cases. Popular applications — such as chatbots which can reduce workloads and lead to faster response times for customers — are having a major impact in the customer service industry. However, these advanced models also face scrutiny by the security community in anticipation of use by attackers.
At Tenable Research, we take seriously the concerns of the broader security community — and we also see enormous potential for the use of LLMs to aid in both offensive and defensive cybersecurity. Our researchers have embraced these technologies, examining how LLMs can be used in both offensive and defensive capabilities, and we’re pleased to share the results of our work in this report.