What a party knows: Rethinking knowledge assessment in automated contracting

In contract law, certain rules require a party’s knowledge of specific facts to operate, such as in cases of unilateral mistake. A mistaken party may be entitled to avoid a contract if their mistake regarding the terms of the contract was known to the other party. However, challenges arise when the other party uses an algorithmic system to automate the contract formation process. In such cases, the party neither directly participates in nor has knowledge of the formation of the contract. This raises a challenge: can the knowledge requirement still be satisfied in the context of automated contracting? This paper argues that in automated contracting, it is difficult, if not impossible, to establish that the party has actual knowledge of the relevant mistake. By contrast, constructive knowledge offers a more viable and conceptually coherent pathway for addressing the party’s absence from the contract formation process. It is proposed that the knowledge of the circumstances surrounding a contract formed by an algorithmic system should be attributed to the party deploying that system. Such attribution is necessary not only to uphold the objective theory of contract formation but is also normatively justified. Under this framework, courts can assess constructive knowledge in automated contracting in the same way as in traditional human-to-human interactions.