Maximizing Representative Efficacy: Part I

Richard Craswell’s excellent “Taking Information Seriously: Misrepresentation and Nondisclosure in Contract Law and Elsewhere” argues our legal system, particularly in the realm of contract law, has fallen short of properly dealing with the “practical issues involved in regulating information.” This manifests in, for example, an inefficient ranking and filtering of information. His article presents significant points for consideration relative to the AiCE project and will be explored here in more detail in subsequent posts.

Agents/entities that fit within the AiCE paradigm are envisioned to perform as wide of an array of legal tasks as the depth of their sophistication is in handling them and, at least initially, as allowed by their designers. And, if we focus solely on the task of “regulating information” we can glean that this will be but a single, broad conceptual category of capability that AiCE can technologically be charged with.

Let us now turn and consider for a moment the setting of a purchase agreement in a B2C environment. Within the class of difficulties classified in the issue of efficiently regulating information festers the question of evaluating whether what initially appears as a misrepresentative statement about a Product (as it is defined in the agreement) is in fact something that necessitates the human party’s attention. Whether this is the case or not is, among other things, dependent on specific deal circumstances, whether or not the human party cares and, if she does, the practical extent that she has the ability to take action to efficiently resolve it.

This is where the concept and capability of environmental learning (aka “adaptive” learning) kicks in. It endows learning capabilities in the AiCE, enabling it to evaluate whether or not a misrepresentative statement in contract A ranks as sufficiently innocuous (within the traditional economic theory of cost-benefit analysis), whereas, in contrast, in contract B the same statement does not resolve as being “true.”

On the most basic of operational levels, (thanks to the AiCE) in either contract setting, the human party is presented with “perfect” information. Of course, that status in and of itself does not justify the means of attaining it. More importantly is what the human party can or cannot do efficiently with that so-called perfect information.

This particular inquiry is relevant to Craswell’s cost-benefit analysis of whether a misrepresentative statement is or isn’t transactionally innocuous. The significantly enhanced computational capabilities that AiCE is envisioned to be capable of dissolves a significant volume of Craswell’s concern and challenge of effectively regulating information.