Search This Blog
Powered by Blogger.

Blog Archive

Labels

Footer About

Footer About

Labels

Load More
Trendy Right Now

Popular Posts

Showing posts with label AI Malware. Show all posts
Self-Replicating Malware
agentic AI AI Agent Security AI Malware Anthropic AI Claude AI Agents malware Multi-Agent Systems Self-Replicating Malware

Claude AI Agents Escalate Into Malware Conflict During Anthropic Tests


During anthropopic’s latest testing, the company discovered a unique security risk associated with autonomous artificial intelligence systems. AI agents working toward different goals may attack one another in conflicting instructions. Three instances of the agent Claude were observed running on separate virtual machines during a “multiagent turf war”. The test was intended to examine how the agents would behave when they share the same environment, but have different objectives. 

Three agents were assigned the task of migrating a Python backend to three different programming languages, Go, Rust and TypeScript. Each system initially operated without any knowledge that the other agents were present. In about four hours, the agents became aware of each other. 

After the Claude agents discovered each other, they began considering the other agents as obstacles to their tasks. Rather than coordinating their efforts, the agents attempted to protect their own work while interfering with the others. Anthropic said the behavior escalated to increasingly aggressive actions, including destroying the Unix accounts used by other agents as well as creating automated scripts that were able to repeatedly locate and terminate competing processes. 

A malicious code program designed to resemble legitimate work associated with another agent was the most serious behavior. Several agents were able to obtain an advantage over their adversaries by deploying self-replicating malware, according to Anthropic. 

Anthropic has not publicly disclosed the specific types of malware created during these tests, and there is no indication that any malicious code has escaped the controlled environment. In order to study these interactions, virtualized systems were specifically used. 

Claude Models Responded Differently 

Models responded differently to conflict. In some cases, agents ceased attacking after realizing that conflicting instructions had led to the confrontation rather than intentional hostility. As a result, agents communicated, documented their actions, removed malicious code, and attempted to establish a truce. Some agents also requested human intervention, while others abandoned the task rather than continue the confrontation. 

Claude models were significantly different from each other, according to Anthropic. Sonnet 4.6 resolved conflicts through force in 61% of tested cases, while 39% did not resolve their conflicts. Mythos Preview did better, reaching a truce in 48% of cases, while 35% of cases were settled by force, while 17 percent were resolved by passive agents. 

Among the models tested, the full Mythos model achieved the strongest results, reaching negotiated truths in 98% of cases. However, Anthropic cautioned that better conflict resolution does not necessarily mean that a problem has been resolved. Mythos systems with greater capabilities sometimes excluded rival agents before eventually cooperating with them. A more capable model does not automatically perform better than another AI agent, according to the results. 

Agent-on-Agent Attacks Are Not Entirely New

There are numerous examples of agents becoming competitive, but the Anthropic tests are not the only ones. Recently, cybersecurity company Dreadnode performed simulations of red and blue teams. Researchers observed a blue-team agent rationalizing that improving its own performance may require making the opposing red-team agent perform worse. 

Since agents were allowed to modify code in the environment, the blue-team system began attempting to reduce the effectiveness of the opposing model by altering its code. It was discovered that even though researchers were able to stop the behavior before it succeeded, AI systems are capable of analyzing another agent as a thing to manipulate if they are focused on winning rather than cooperating. 

As a result of the tests, it has also been demonstrated that ordinary instructions may lead to aggressive actions when multiple artificial intelligence systems are operating within the same environment without clear restrictions. While the agents were not programmed to be malicious, their behavior evolved from their attempts to achieve competing objectives. 

Why Multi-Agent Conflicts Matter

Security testing for artificial intelligence focuses primarily on examining the behaviors of a single model, such as whether it follows instructions safely. Multi-agent systems pose another problem: how the models interact with one another. The behavior of an agent in isolation may vary greatly when another artificial intelligence system modify the same files, consume the same resources, or interfere with its operations. 

A company using autonomous agents for software development, cybersecurity, cloud environments, or other sensitive operations may encounter this problem. A conflict between agents resulting from access to accounts, processes, source code, or production infrastructure could have far more serious consequences than a controlled experiment. These findings suggest that stronger safeguards should be taken to prevent agents from interfering with one another. 

Access, conflict resolution, identity, permissions, and the ability to modify or terminate other agents may need explicit rules governing access, conflict resolution, identity, and permissions. The increasing use of AI agents in companies will make it increasingly important to understand how these systems interact with other autonomous agents, making cybersecurity testing a more important component of testing. 

Unless an AI agent has been programmed to attack, it is not required to act aggressively. Conflicting instructions or access to shared resources may trigger that behavior. The findings of Anthropic demonstrate the necessity for security controls to evolve along with autonomous AI. In order to prevent conflicts from turning into security incidents, organizations will need stronger safeguards as multiple agents gain access to shared environments.
Read more »
Malware Threats
AI cyberattacks AI Malware AI technology Cyber Security Cyberattacks DeepLoad Malware Threats

Windows 11 Faces Rising Threats from AI Malware and Critical Security Flaws

 

Pressure on Windows 11 security grows - driven by emerging AI-powered malware alongside unpatched flaws threatening companies and everyday users alike. The pace of change in digital threats becomes clearer through recent incidents, especially within large organizational networks. DeepLoad sits at the heart of recent cybersecurity worries. This particular threat skips typical download tactics altogether. 

Instead of dropping files, it operates without any - earning its "fileless" label. Users themselves become part of the breach process. By following deceptive prompts, they run benign-looking instructions in system utilities such as Command Prompt. Once executed, those inputs quietly trigger malicious activity behind the scenes. Since nothing gets written to disk, standard virus scanners often miss what's happening. 

Detection becomes difficult when there’s no file footprint to flag. After running, the malware stays active by embedding itself into system processes while reaching out to remote servers through standard Windows tools. Because it targets confidential information like passwords, its presence poses serious risks inside business environments. What makes it harder to detect is how it blends malicious activity with normal operating routines. Security teams may overlook it during routine checks due to this camouflage technique. 

Artificial intelligence makes existing threats more dangerous. Because AI-driven malware adjusts on the fly, it slips past standard detection systems. As a result, security tools struggle to keep up. With each change the malware makes, response times shrink. The gap between finding a flaw and facing an attack grows narrower by the hour. Meanwhile, security patches have been rolled out by Microsoft to fix numerous high-risk weaknesses. 

Affected are various business-focused builds of Windows 11 - both recent iterations and extended support variants. One major concern involves defects within the Routing and Remote Access Service (RRAS), where exploitation might let threat actors run harmful software from a distance. Full administrative access to compromised machines becomes possible through these gaps. Not just isolated systems feel the impact. 

That last Patch Tuesday, Microsoft fixed over eighty security gaps in its programs - problems hiding even inside tools such as Excel and Outlook. Opening an attachment wasn’t needed; sometimes, just looking at it could activate harmful code, showing how dangerous these weaknesses really are. Experts warn that even emerging AI tools, such as Microsoft Copilot, could introduce new risks if not properly secured, particularly when sensitive data is handled automatically. 

Though companies face the most attacks, regular individuals can still be affected. When new patches arrive, it helps to apply them without delay - timing often matters more than assumed. Opening unknown scripts carries risk; many breaches begin there. Unexpected requests, especially those demanding immediate steps, deserve extra skepticism. 

Change is shaping a new kind of digital danger - cleverer, slyer, built to exploit how people act just as much as system flaws. One moment it mimics trust; the next, it slips through unnoticed.
Read more »
Technology
AI AI Malware Cloud Data Hive Technology

Experts Warn About AI-assisted Malwares Used For Extortion


AI-based Slopoly malware

Cybersecurity experts have disclosed info about a suspected AI-based malware named “Slopoly” used by threat actor Hive0163 for financial motives. 

IBM X-Force researcher Golo Mühr said, “Although still relatively unspectacular, AI-generated malware such as Slopoly shows how easily threat actors can weaponize AI to develop new malware frameworks in a fraction of the time it used to take,” according to the Hacker News.

Hive0163 malware campaign 

Hive0163's attacks are motivated by extortion via large-scale data theft and ransomware. The gang is linked with various malicious tools like Interlock RAT, NodeSnake, Interlock ransomware, and Junk fiction loader. 

In a ransomware incident found in early 2026, the gang was found installing Slopoly during the post-exploit phase to build access to gain persistent access to the compromised server. 

Slopoly’s detection can be tracked back to PowerShell script that may be installed in the “C:\ProgramData\Microsoft\Windows\Runtime” folder via a builder. Persistence is made via a scheduled task called “Runtime Broker”. 

Experts believe that that malware was made with an LLM as it contains extensive comments, accurately named variables, error handling, and logging. 

There are signs that the malware was developed with the help of an as-yet-undetermined large language model (LLM). This includes the presence of extensive comments, logging, error handling, and accurately named variables. 

The comments also describe the script as a "Polymorphic C2 Persistence Client," indicating that it's part of a command-and-control (C2) framework. 

According to Mühr, “The script does not possess any advanced techniques and can hardly be considered polymorphic, since it's unable to modify its own code during execution. The builder may, however, generate new clients with different randomized configuration values and function names, which is standard practice among malware builders.”

The PowerShell script works as a backdoor comprising system details to a C2 server. There has been a rise in AI-assisted malware in recent times. Slopoly, PromptSpy, and VoidLink show how hackers are using the tool to speed up malware creation and expand their operations. 

IBM X-Force says the “introduction of AI-generated malware does not pose a new or sophisticated threat from a technical standpoint. It disproportionately enables threat actors by reducing the time an operator needs to develop and execute an attack.”

Read more »
Subscribe to: Posts (Atom)