German retailers see tenfold rise in unauthorized AI use, prompting security alerts

A new Omnissa study shows that about one‑third of stationary devices in German retail stores run operating systems that are four to five generations old, causing checkout delays and inefficient inventory processes. At the same time, employee use of unauthorised AI tools such as ChatGPT and Gemini has increased almost tenfold year over year, creating shadow‑IT, compliance gaps and heightened risk of data breaches. Recent sophisticated cyber‑attacks on retailers including Tegut, Media Markt, KaDeWe and Ceconomy have exposed thousands of customer records and disrupted server infrastructure.

Security experts warn that simply inventorying AI agents is insufficient. Effective protection requires controlling identity, purpose and context for each agent and enforcing least‑privilege access, especially as AI agents now operate autonomously in SaaS, cloud‑workflow and productivity tools. Real‑time transparency, automated update mechanisms and robust internal controls are deemed essential to mitigate the growing threat landscape.