A study combining an audit of 100 start-ups with 63 interviews with experts shows how the ‘AI’ label is overhyped to attract investors, and how it subsequently erodes consumer confidence.
AI-washing refers to the use of the word AI as a marketing argument beyond what the product actually does. Researchers screened 100 German start-ups using a seven-dimensional annotation grid and conducted 63 interviews with experts in a study published on arXiv on 8 October.
Three Ways to Exaggerate
Exaggerations take three forms: statements that cannot be verified, performance figures without methodology, and promises presented without limits. The boasted capabilities rarely correspond to the underlying system. And the effect is paid later: consumers already exposed to swollen promises withdraw their confidence in the following AI communications. The authors also report a link to problematic uses, with 38% of the sample raising moderate or serious ethical concerns, including monitoring and automated decision-making in regulated sectors.
The mechanism described is simple: the founders inflate the capacity to attract investors who reward an AI label they do not know how to verify, while customers, unfamiliar with the domain, take the speech at the foot of the letter. The team proposes to subject AI promises from start-ups to transparency obligations, and makes its grid available to regulators, investors and consumer associations for regular audits.
Read the full article on the original source
Go to the source →