AI-driven hackers breach thousands of business accounts via software flaw

A cyberattack campaign led by a China-aligned group has successfully exploited a vulnerability in a widely used business management platform to gain unauthorized access to over 2,250 Microsoft 365 tenants. The incident highlights a new phase in digital warfare where artificial intelligence accelerates the speed and scale of intrusions, leaving corporate defenders with very little time to react before attackers reach the highest levels of system privilege.

The primary claim of this event is that the use of open-source AI agents fundamentally changed the dynamics of this supply chain attack. The group, identified by Microsoft as Storm-0940, utilized ARTEX, a framework that allows large language model driven agents to perform autonomous penetration testing. This technology enabled the hackers to conduct reconnaissance, scan for weaknesses, and execute password guessing campaigns at a pace that traditional human teams could not match. The agents operated continuously with minimal human oversight, compressing the timeline from initial access to full control into a matter of hours.

The mechanics of the intrusion relied on stolen credentials for ConnectWise applications. These credentials allowed the attackers to authenticate as the applications themselves, bypassing the need for interactive user sign-ins. By leveraging these app identities, the AI agents were able to sign in to customer tenants from the same IP range. In one rapid sequence, more than 200 sign-ins were recorded in under four minutes using just four application credentials. This automated sweep indicated a scripted approach working from a pre-defined target list, moving through tenants in sorted order with only seconds between attempts.

The speed of the attack was particularly alarming. Some of the intrusions reached privileged access less than three hours after the initial targeting began. On the peak day of the password spraying component, which targeted 327 accounts across eight organizations, the system attempted nearly 111,000 logons in a single day. While almost 98% of these automated guesses resulted in lockouts, the success of the application-based sign-ins allowed the attackers to perform directory operations, change app credentials, and modify user roles. In one specific case on October 15, an app added itself to the top administrator role, a behavior that triggered specific security alerts.

However, there are significant caveats regarding the extent of the damage. It remains uncertain how the initial ConnectWise credentials were stolen or exactly how many tenants suffered data exfiltration. Microsoft clarified that the attack did not stem from a flaw in its own systems or products. The presence of an affected app or a successful sign-in does not automatically mean that data was accessed or taken. Security firm Huntress confirmed unauthorized sign-ins for 162 tenants by October 16 and nearly 1,000 by the following day, with more than 300 of those sign-ins succeeding. This suggests a broad but potentially shallow compromise in many cases.

The choice of targets revealed the strategic intent behind the operation. The group focused on high-value industries including aerospace and defense, semiconductors, government agencies, and think tanks. Fourteen specific industries were listed as targets in the analysis. Storm-0940, which has been active since 2021, is described as a developing actor with no known industry aliases. The use of AI in this context is not isolated to this group, as public reports this month linked similar breaches at South Korean financial institutions to the same technology.

The implications for the broader tech industry are significant given the status of the tool used. ARTEX, which can utilize any AI model, has been adopted by threat actors in multiple regions. Notably, the developer of ARTEX, known by the handle Autumn-27, announced on GitHub that the project would no longer be updated and would become closed source. This decision may alter how future attackers access the framework, but the immediate impact of its use in this campaign is clear.

ConnectWise has suspended several of its cloud services, including Cloud Backup and SaaS Security, as a precautionary measure. The company has begun a phased restoration process, with recovery expected to take several days. There are currently no indicators that partner backup data was compromised. Microsoft advises customers to keep affected app connections disabled until formal restoration instructions are issued and to review recent activity logs where possible. This incident serves as a stark reminder that the integration of AI into offensive cyber operations is rapidly outpacing traditional defensive measures.