Research Summary: Auditing AI-Washing in German Startups: A Mixed-Methods Study of Marketing Claims and Perceptions

A mixed-methods empirical study auditing 100 German startups and incorporating 63 expert interviews identified widespread AI-washing, wherein marketing claims systematically exaggerated underlying technological capabilities. The study revealed that startups frequently present unverified claims, ungrounded performance figures, and omit operational limitations, leading to an erosion of consumer trust across subsequent artificial intelligence marketing.

Why it matters

This research matters because systemic overstatement of artificial intelligence capabilities creates significant reputational risk and market skepticism. Widespread AI-washing erodes stakeholder trust and misleads investors, ultimately threatening the perceived legitimacy of genuine technological innovation across the startup ecosystem.

Key insights

• Startups routinely exaggerate artificial intelligence capabilities, with claimed features rarely matching underlying technical systems.
• Exaggeration manifests primarily in three forms: claims lacking inspectable evidence, performance metrics presented without explanatory methodology, and claims omitting technical limits.
• Consumer exposure to exaggerated claims results in reduced trust toward subsequent AI-related marketing communications.
• The study examined 100 German startups across a seven-dimension annotation codebook combined with insights from 63 expert interviews.

Source

arXiv — Computers and Society — https://arxiv.org/abs/2610.11788