Hackers Developing ChatGPT-like AI for Scams; Stay Alert

AI Phishing Scam: Hackers are constantly evolving alongside the rise of AI. Reports now indicate that hackers are creating Large Language Models (LLMs) similar to ChatGPT to facilitate cyberattacks.

AI Phishing Scam: Cybercrime has surged with the advent of AI. Cybercriminals are refining their methods using AI tools, making it increasingly difficult for people to avoid falling victim to them. A new report has now raised concerns regarding cybersecurity. It reveals that a North Korea-linked hacking group is developing an LLM akin to ChatGPT; this model is intended to analyze information, automate tasks, and devise more sophisticated scams. Simply put, hackers are building a new AI model specifically to carry out scams. Let’s explore this in detail.

What is an LLM?

A Large Language Model (LLM) is an advanced AI system built on deep neural networks, designed to process and understand data and generate human-like text. These models answer questions by grasping patterns, grammar, and context from text. They can perform various tasks, such as content creation and language translation. LLMs are primarily built using transformer architecture, which enables them to comprehend the meaning and context of text. Key benefits include the ability to easily understand input data and automate repetitive language-based tasks.

Hackers Developing Software for LLMs

According to media reports, the South Korean cybersecurity firm Genians has found evidence that the hacking group Kimsuky is developing software to run LLMs on its own systems. The group is utilizing tools such as Ollama, GPT4All, and Msty, which facilitate the local operation of AI models. These tools eliminate the need for models to rely on cloud-based services. Researchers discovered that hacking groups are utilizing Retrieval-Augmented Generation (RAG) technology, which enables AI systems to search through and comprehend collections of documents. By leveraging these tools, hackers can analyze sensitive files without relying on external platforms.

Rising Threats in the Future!

Cybercriminals are already using AI tools to craft phishing emails that are extremely difficult to detect. Now, by adopting new methods, hackers are planning even more sophisticated attacks. Genians reported the discovery of AI-assisted coding, speech-to-text capabilities, and the development of AI agents. These tools allow hackers to automate their malicious activities. Furthermore, they can rapidly analyze stolen data and potentially even create malware. Researchers note that hackers are no longer limiting AI usage to fake messages; instead, they are integrating AI into various stages of cyberattacks. This has heightened the risk of increasingly advanced cyberattacks in the future.

AI-Generated Fake Documents

The Genians report also highlights numerous AI-generated documents related to finance and cryptocurrency. These documents appear completely authentic and seem to originate from legitimate companies. Hackers use them to lure individuals into opening malicious files and sharing sensitive information. Therefore, it is crucial to remember that an email or message appearing genuine does not necessarily come from a legitimate company; it could be a tactic used by hackers to ensnare you. In this era of AI, prioritizing online safety is essential. Always verify the sender before opening any email. If you receive an email from a suspicious or unknown sender, do not open it.