AI against criminals: ChatGPT now investigates cybercrime

Cybersecurity researchers from several universities appreciated the impact of ChatGPT especially GPT-4for digital forensics.

Rapid progress in the development of large language models (Large Language Model, LLMs), including GPT-3.5, GPT-4, BERT, Bard and others, opens up new horizons for the application of AI in the field of digital forensics. Today, the capabilities of artificial intelligence can be used to:

• data classification;
• network forensics;
• malware analysis;
• face recognition systems.

However, despite the rich potential of these technologies, they can make mistakes, which raises legal, ethical and research questions of excessive dependence on technologies. In addition, technical limitations should also be taken into account.

The scientists emphasized the importance of being able to quickly create scripts in a resource-limited environment, which is critical for solving problems such as file recovery, working with RAID arrays (Redundant Array of Independent Disks – a redundant array of independent (independent) disks), cracking passwords and recovering encryption keys.

The authors of the article note that despite all the obvious advantages of using ChatGPT in the field of digital forensics, AI has a number of functionality limitations. The study was not exhaustive, and the actual complexity may differ significantly from the conditions of the experiments. In addition, the results largely depend on the tasks set and may not be fully reproducible due to the non-deterministic nature of ChatGPT.

In spite of all this, AI continues to develop actively in the field of cybersecurity, as evidenced by the emergence of the tool WormGPT – analogue of ChatGPT for pentest.