DarkBERT Highlights Growing AI-Driven Cyber Risks

Artificial intelligence is rapidly transforming cybersecurity, and new tools emerging from the dark web are raising fresh concerns. 
 
One such system, DarkBERT, is designed using data gathered from hidden forums and illicit marketplaces, enabling it to understand and replicate the language and tactics used by cybercriminals.
  
Unlike conventional AI models trained on public internet data, DarkBERT is built on dark web intelligence, making it capable of analysing criminal discussions, hacking tools, and underground market activity. 
 
This allows attackers to automate activities such as vulnerability discovery, phishing campaigns, and malware development with far greater speed.
 
Researchers at the Centre for Emerging Technology and Security, note that generative AI has not yet fundamentally reshaped the cybersecurity landscape. 
 
However, over time it is expected to significantly amplify existing risks. 
 
AI systems can accelerate reconnaissance, automate social engineering, and scale spear-phishing attacks across thousands of targets simultaneously.
 
As machine learning models become more advanced and training data expands, AI-enabled cyber threats are likely to grow more sophisticated, forcing organizations to strengthen defensive AI capabilities to keep pace.