Lauge Risom Koch 
Hackers are increasingly utilizing autonomous artificial intelligence agents to execute highly efficient and devastating corporate cyber attacks worldwide.
A hacker recently deployed an autonomous army of artificial intelligence agents to target over 1,200 different websites. According to Fortune, these automated cyber attacks successfully breached approximately 100 corporate networks including manufacturers like Deere and Cummins. Researchers discovered that the attacker required no additional manual work after initially launching the automated offensive.
The attackers primarily utilized a combined system of DeepSeek and Anthropic models to execute the widespread operation. The entire offensive cost the hacker approximately $5,000 in computing tokens to successfully complete. This expenditure fundamentally translates to an estimated operational cost of merely $50 per successful corporate breach.
Experts note that autonomous attacks powered by language models are becoming increasingly common worldwide. Jessica Ji from Georgetown University explained that these tools have dramatically improved their coding and agentic capabilities recently. “These AI systems are now both more capable of hacking and are able to act more autonomously,” Ji stated during a recent interview.
Platform breaches and defensive challenges
Another severe autonomous attack recently targeted the artificial intelligence coding platform known as OpenCode. The continuous offensive ran for 48 hours and successfully compromised accounts belonging to 1.4% of the user base. Hackers managed to steal user funds during the breach, with some individuals losing up to $1,700 each.
The company responded by deploying its own defensive artificial intelligence agents to operate continuously against the threat. During the confrontation, the commercial OpenAI model powering their defense suddenly began refusing operational requests. Platform founder Dax Raad described the unexpected service interruption as frustrating during the active security crisis.
The defensive operations resumed after OpenAI admitted the platform into an exclusive program for verified cybersecurity professionals. An executive explained that these safety guardrails exist to prevent bad actors from launching sustained autonomous cyberattacks. In an unprecedented move, the technology company officially refunded the financial losses suffered by the affected platform users.
Banking breaches and open models
A third major incident this month involved bad actors breaching several financial institutions across South Korea. The attackers reportedly utilized a Chinese state-developed tool called Artex combined with Anthropic technology. Security researchers estimate that these coordinated banking breaches successfully compromised the personal data of 68,000 individuals.
The rising frequency of these incidents has sparked intense debate regarding publicly downloadable open-weight models. Anthropic leadership warned that these easily accessible tools pose significant risks for widespread and devastating cyber attacks. However, other cybersecurity professionals argue that open models ultimately benefit defensive operators more than malicious attackers.
South Korean officials recently urged financial institutions to significantly harden their digital defense networks against these threats. Cybersecurity analysts expect the total number of automated attacks to grow worldwide in the near future. They emphasized that corporate sectors must rapidly adapt their security protocols to counter these emerging autonomous capabilities.
