-
Views
-
Cite
Cite
Philipp Hacker, Personal data, exploitative contracts, and algorithmic fairness: autonomous vehicles meet the internet of things, International Data Privacy Law, Volume 7, Issue 4, November 2017, Pages 266–286, https://doi.org/10.1093/idpl/ipx014
- Share Icon Share
Extract
Introduction
...
From health care to party games, specialized Artificial Intelligence (AI) is increasingly deployed to automate tasks that require differentiated pattern recognition and decision-making previously thought to be reserved to humans alone.1 Simultaneously, the Internet of Things (IoT) is connecting devices and sensors into an ever more intricate and encompassing net of data collection and analysis.2 However, these developments are increasingly intertwined:3 epitomized by the advent of autonomous and connected vehicles (ACVs), the merger of AI with the IoT holds a vast potential to reshape the boundaries of public and private spheres, and to transform human interactions with their immediate environment, for better or worse.This article analyses one particularly pressing threat that arises from this merger. Personal data collected via the IoT, and processed with the power of specialized AI, may be used for a specific type of strategic behaviour that is increasingly investigated in the economics literature, but largely neglected by the legal debate in this context: exploitative contracts.4 Such contracts are tailored to vulnerabilities of counterparties that become apparent in data patterns. In the economic literature, a contract is considered exploitative if ‘the economically central considerations driving it derive from trying to profit from the agent’s mistake’.5 Similarly, in legal analysis, contracts are deemed exploitative if a party consciously takes advantage of a position of strategic advantage vis-à-vis the counterparty to influence the formation of a contract.6 Such contracts not only materially disadvantage the exploited party, but also contradict notions of autonomy7 and fairness8 inherent in contractual dealing. With AI technologies increasingly deployed in marketing and customer analysis,9 the recognition of vulnerabilities will be further facilitated, and exploitative opportunities are likely to grow in the future. A leading textbook on machine learning notes that contractual offers ‘provided specifically to retain existing customers can be expensive, and successful data mining allows them to be precisely targeted to those customers who are likely to yield maximum benefit’;10 when companies have ‘complete purchasing histories for each individual customer, [they] can use data mining to determine those likely to respond to special offers’.11 Such targeted offers can be mutually beneficial, but they can also be exploitative if they are valuable for the company only, and not for the target. A key challenge for the law, hence, is to accommodate this ambivalence, and to rein in data-driven exploitative contracting.