The Federal Office for Information Security and the TÜV Association have presented a comprehensive assessment of the evolving cyber threat landscape across the German economy. The findings stem from a structured forsa survey encompassing 505 companies with at least 20 employees across multiple industries. In that survey, 17 percent of businesses reported being targeted by attacks or concrete fraud attempts involving artificial intelligence over the past twelve months. These figures demonstrate that machine-assisted attack patterns have definitively moved beyond theoretical scenarios to become a tangible part of day-to-day operations.
A closer examination of the reported security incidents reveals the severe evidentiary difficulties that corporate security personnel currently face. Exactly four percent of surveyed enterprises verified confirmed security breaches with unambiguous AI involvement, while eleven percent recorded concrete suspected cases. An additional two percent of participants reported experiencing a combination of confirmed attacks alongside unverified suspicious activity. This distribution points to a substantial number of undetected incidents, as sophisticated attacks often go unnoticed or fail to be identified as machine-generated.
When looking at specific attack vectors, communication channels that can be convincingly manipulated using generative technologies clearly dominate. Fully 90 percent of affected businesses encountered highly personalized phishing emails that mimic nuanced language conventions and easily bypass traditional spam filters. Furthermore, 40 percent of targeted organizations recorded the deployment of automated attack scripts designed to rapidly scan for and exploit network vulnerabilities. In eleven percent of cases, criminals turned to deepfakes, generating synthetic audio or video impersonations of corporate executives to execute CEO fraud schemes.
Despite this transformed threat environment, the survey exposes critical organizational deficits across corporate risk management practices. Only eleven percent of companies that already use or plan to introduce artificial intelligence maintain formal incident response processes tailored to AI-related attacks. Most organizations lack binding escalation procedures to react swiftly and systematically to synthetic identity fraud or automated intrusions. Administrative preparation has therefore failed to keep pace with the swift technological advancements seen on the attacker side.
The situation remains equally concerning on the technical defense front, exposing a clear asymmetry between attackers and defenders. Merely 31 percent of companies actively using artificial intelligence deploy defensive AI systems to detect and mitigate incoming threats. The remaining majority continues to rely strictly on conventional security tools that are rarely equipped to match the speed and adaptability of algorithmically driven incursions. Without machine-assisted defensive capabilities, internal security teams face severe disadvantages in containing automated attacks in real time.
The findings from this joint investigation underscore the imperative to adapt security frameworks to rapid technological shifts. Basic staff training focused on identifying traditional warning signs in malicious messages is no longer adequate when generative models produce flawless, tailored correspondence in seconds. For enterprises, establishing structured governance protocols, standardized reporting lines, and automated defensive systems has become an operational necessity. Only when robust organizational planning works in tandem with modern defense tools can businesses effectively curtail the risk of severe financial and operational damage.