ECLI:NL:GHDHA:2024:1771

ChatGPT ai assisted selection of comparatorsoverreliance on ai outputinsufficient substantiation of comparability

I. Executive Summary

In a bpm dispute about determining comparable vehicles, the appellant used ChatGPT to generate a list of cars said to be in the same economic context and competitive position. The court held that such an AI-generated list, without further substantiation, is insufficient to establish which vehicles are the most similar; similarity must be assessed by reference to what an average consumer (a human) would regard as similar. The appeal was dismissed.

II. Conduct Analysis

A party used ChatGPT to produce a list of purportedly comparable vehicles and relied on that output to argue similarity and valuation, without demonstrating comparability by established legal criteria.

III. Legal Foundations

Wet bpm articles 9 and 10 article 110 TFEU CJEU comparability approach (average consumer)

IV. Key Facts

1) The appellant used ChatGPT with a prompt asking for as many cars as possible in the same economic context and competitive position. 2) ChatGPT produced a list of ten exclusive cars, which the appellant relied on as comparators. 3) The court held that more is required than an AI program stating competitive context; the reference point is what the average consumer would consider similar.

V. Consequences & Sanction

1) The court rejected reliance on the ChatGPT-generated comparator list as a basis for legal comparability. 2) The appeal was dismissed.