Kelly-Brown Ochieng Odhiambo (Umma University) has posted The Automaton as Intervening Cause: Algorithmic Amplification, Causal Attribution and the Quantum Conundrum in Kenyan Digital Defamation on SSRN. Here is the abstract:
Digital publication has fractured an assumption upon which much of defamation law has historically operated: that the person responsible for communicating defamatory matter is also substantially responsible for determining the audience to which it travels. Algorithmic recommendation, ranking, and personalisation systems separate those functions. A media organisation may originate defamatory material, while an independent platform determines whether, when, and to whom that material is subsequently presented. The resulting problem is not simply one of intermediary liability; it is a fundamental problem of causation and remedial attribution.
Kenyan defamation law presently possesses no developed doctrine addressing the causal consequences of autonomous distribution. Its existing rules concerning publication, material contribution, novus actus interveniens, remoteness, republication, and damages nevertheless provide the conceptual materials from which a distribution-sensitive account of the tort can be developed. The central proposition advanced here is that algorithmic amplification should be treated neither as legally irrelevant automation nor as an autonomous tortfeasor. It is a distinct causal mechanism whose legal significance depends upon the structural relationship between the originating defendant, the distributional system, and the resulting injury.
The article develops this proposition through a jurisprudential distinction between cause in-fact and scope of liability. Drawing upon Hart and Honoré, Wright, and Stapleton, it argues that an originating publication and an independent amplification event can constitute distinct causal contributions without requiring either to be characterised as the exclusive cause of the claimant’s injury. Drawing upon corrective-justice theory, it further argues that attribution must remain correlative: the defendant should answer for the consequences of the wrong attributable to that defendant, not for an aggregate digital injury simply because another responsible actor is immune or unavailable.
The article articulates the concept of algorithmic causal independence as a factual condition in which an automated distribution mechanism is sufficiently independent of the originating publisher, and sufficiently material to the subsequent injury, to require separate causal attribution. This framework resolves the ongoing attribution-quantum disconnect in digital actions, provides a structured methodology for dividing damages between original platforms and viral systems, and proposes a model for statutory reform through a newly drafted Section 16B of the Defamation Act (Cap 36).
Recommended!
To receive new posts from Legal Theory Blog by email, get a free subscription to Legal Theory Stack.
Lawrence Solum