AI-infused contracting and the problem of relationality: Is trustworthy AI possible?

Activity: Talk or presentationWorkshop

Documents

TT Arvind - Invited speaker

The focus of this paper is on a relatively new, but rapidly growing, phenomenon which I term 'AI-infused contracting'. AI-infused contracting holds considerable promise for the overall quality and effectiveness of transactions in the commercial and non-commercial worlds. Nevertheless, this paper argues that its rise is not problem-free. Making the AI systems that underpin it resilient and trustworthy involves non-trivial challenges, which go to the heart of the nature of contract law and contract practice in the present day. AI in the real world is not simply a technical system, but a socio-technical system. Its real world use depends not just on the technology underpinning it, but necessarily also on the manner in which and ends towards which that technology is deployed, and on the predispositions, asymmetries, and biases that characterise the specific social contexts in which those ends are pursued. The argument of this paper is that this makes AI something of a double-edged sword in the domain of contracts. In each of the areas in which it is used, AI has the potential to transform the drafting, management, and implementation of contracts in a way that ameliorates a significant proportion of the problems posed by existing contracting practices. But, equally, it has the potential to materially exacerbate these problems, in part because of the nature of AI itself and more specifically of the heuristics that underpin it, and in part because of the limits of law's regulatory capacity. Addressing this risk, and creating AI that is trustworthy in relation to contracting, will require channeling its use in a new direction, which this paper describes as 'transactional responsibility'. Achieving and implementing this standard will, in turn, require rethinking not just contract law but also general programming practices and approaches in the field of contract-related AI.
19 Jul 202123 Jul 2021

External organisation

NameNational University of Singapore
CitySingapore
CountrySingapore

Projects

Discover related content

Find related publications, people, projects, datasets and more using interactive charts.

View graph of relations