A recent study analyzing over 175,000 posts from cybercrime forums demonstrates that artificial intelligence has not made hackers more dangerous but has instead made their code notably sloppier. Contrary to fears that AI would empower cybercriminals with advanced skills and methods, the research shows that AI tools are often used to cut corners rather than to deepen expertise.
The study highlights that cybercriminals adopt AI similarly to everyday users—primarily to avoid learning the underlying details. Hackers increasingly rely on AI outputs without thoroughly reviewing or refining the generated code, leading to more errors and vulnerabilities within their own malicious software. This casual approach undermines the presumed AI-driven leap in hacker capability.
Researchers examined a broad sample of online discussions and shared code snippets, observing frequent mentions of AI tools assisting with script generation, phishing message crafting, and malware development. Despite this, the quality of AI-assisted hacking tools tends to degrade, as cybercriminals often accept ready-made solutions without adapting them adequately.
The findings suggest the integration of AI into cybercriminal activities represents convenience rather than innovation, with adversaries prioritizing speed over precision. While AI can automate complex tasks, in the hands of hackers, it currently contributes more to sloppy execution than to sophisticated attacks.