GenEth: A General Ethical Dilemma Analyzer

Authors: Michael Anderson, Susan Anderson

AAAI 2014 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental From all six cases of the example domain presented previously, the following disjunctive normal form principle, complete and consistent with respect to its training cases, was abstracted by GENETH:... To evaluate the principles codified by GENETH, we have developed an Ethical Turing Test... Of the 140 questions, the ethicists agreed with the system s judgment on 123 of them or about 88% of the time.
Researcher Affiliation Academia Michael Anderson Susan Leigh Anderson Dept. of Computer Science, U. of Hartford Dept. of Philosophy, U. of Connecticut
Pseudocode No The paper describes the learning algorithm verbally and through examples but does not present it in a structured pseudocode or algorithm block.
Open Source Code Yes (An OSX version of the software is freely available at:http://uhaweb.hartford.edu/anderson/Site/Gen Eth.html)
Open Datasets No The paper uses 'cases' which are input by ethicists ('As new cases of a given ethical dilemma are presented to the system...'). It does not mention or provide access information for a publicly available or open dataset.
Dataset Splits No The paper mentions that the Ethical Turing Test questions were drawn from 'training (60%) and non-training cases (40%)', which refers to the evaluation set, not a specified validation split for the model's training process.
Hardware Specification No The paper does not provide any specific details about the hardware (e.g., GPU/CPU models, memory) used for running the experiments.
Software Dependencies No The paper mentions 'inductive logic programming (ILP)' and 'Allegro Common Lisp s Metaobject Protocol' but does not provide specific version numbers for software dependencies or libraries.
Experiment Setup Yes GENETH starts with a principle that simply states that all actions are equally ethically preferable (that is p(a1,a2) returns true for all pairs of actions). An ethical dilemma and two possible actions are input, defining the domain of the current cases and principle. The system then accepts example cases of this dilemma... GENETH s approach is to incrementally specialize a principle so that it no longer returns true for any negative cases (those in which the second action is deemed preferable to the first) while still returning true for all positive ones (those in which the first action is deemed ethically preferable to the second).